NTT Develops Real-Time Voice Conversion GenAI Tech That Can Instantly Change Your Voice and Speaking Style

NTT Corporation has recently announced that it has developed a remarkable real-time voice conversion technology based on deep learning that achieves both high sound quality and low latency.

This technology enables voice conversion in a variety of voice communications, whether face-to-face or remotely, and contributes to the realization of communication that is free from physical, intellectual, and psychological constraints, for example, converting the intonation and voice quality of a speaker into easy-to-understand speech at a call center.

High Sound Quality and Low Latency: The technology achieves both high sound quality and low latency. Unlike conventional methods, it doesn't require a buffer for future speech signals, resulting in real-time conversion without delays.

Voice Feature Extraction: A newly devised voice feature extraction process ensures high sound quality. It flexibly converts voice quality, intonation, and rhythm using paired data of the same utterance of the source and target speakers.

Applications

This breakthrough enables voice conversion in various scenarios, whether face-to-face or remote. Imagine converting the intonation and voice quality of a speaker into easy-to-understand speech at a call center or during web conferencing.

Communication Enhancement through Voice Conversion

The technology opens doors for web conferencing, live streaming, and smartphone applications. It contributes to communication free from physical, intellectual, and psychological constraints.

This technology is expected to enrich speech communication in various business and real-life situations, whether face-to-face or remote, such as the use of this technology for dysphonia, fluent English pronunciation close to native English, persuasive speech, and removing of nervousness-induced voice tremors, etc.

In the future, NTT says that it will work to improve noise-resistance and stability in real environments, as well as countermeasures against impersonation, with the aim of creating a future in which users can communicate with their favorite voices more securely.

Researchers Introduce AI, Deep Learning Model for Improved Earthquake Forecasting

The day after the massive riots in Minneapolis and Lake Street on May 29th. Here was an O'Reilly Auto Parts that was burnt down, but right above was a billboard that remained intact that says "Coming Soon". A foreboding prediction? [Image - Parker Johnson@Unsplash ] 

Researchers from the Universities of California at Berkeley and Santa Cruz, and the Technical University of Munich recently introduced the Recurrent Earthquake foreCAST (RECAST), a deep learning model for earthquake forecasting and explore its performance on synthetic and regional earthquake data sets.

A typical earthquake forecasting relied heavily on statistical models that do not scale to fully utilize the currently available large earthquake data sets. The researchers build on recent developments in deep learning for forecasting event sequences in general to create an implementation for earthquake data.

In their recently released a paper, the researchers said that the deep learning model can use larger datasets and offer greater flexibility than the current model standard, ETAS, which has improved only incrementally since its development in 1988.

The researchers used NVIDIA GPU workstations to train the model.

The promise of RECAST is that its model flexibility, self-learning capability and ability to scale will enable it to interpret larger datasets and make better predictions during earthquake sequences.

As of now, ETAS model is the most popular model to describe frequency of earthquakes in a region and is based on the theory that every earthquake has a magnitude-dependent ability to trigger aftershocks. The ETAS model considers that the earthquake sequence comprises aftershocks (events triggered by other earthquakes) and background earthquakes (events that occur independent of other earthquakes).

RECAST is fundamentally different from ETAS. The exact functional form that relates catalog data (records of earthquake data) to the likelihood of the next event time is NOT required.

Tests on synthetic data suggest that with a modest-sized data set, RECAST accurately models earthquake-like point processes directly from cataloged data. The research paper further says that — Tests on earthquake catalogs in Southern California indicate improved fit and forecast accuracy compared to our benchmark when the training set is sufficiently long (>10⁴ events). The basic components in RECAST add flexibility and scalability for earthquake forecasting without sacrificing performance.

RECAST builds on recent developments in machine learning known as neural temporal point processes or TPPs. The model uses a general-purpose encoder-decoder neural network architecture that predicts the timing of the next event given the history of past events.

This research study is a proof of concept for the application of neural temporal point process models for earthquake forecasting. More generally, neural temporal point process models and deep learning provide important advantages. They do not require knowledge of the relationship between earthquake features to forecasting probabilities.

With the larger datasets, the researchers are starting to see improvements from RECAST over the standard ETAS model.

This new deep learning based approach allows researchers to incorporate larger data sets, potentially with more information about each earthquake

To advance the state of the art in earthquake forecasting, Dascher-Cousineau, who is one of the authors of the research papers, is working with a team of undergraduates at UC Berkeley to train earthquake catalogs on multiple regions for better predictions.


Via ~ NVIDIA Blog

1st Batch of IIT Ropar and The Second Wind Post Graduate Certificate in Artificial Intelligence & Deep Learning (PGCAIDL) completed

Students of the Post Graduate program in ‘Artificial Intelligence and Deep Learning’ (PGCAIDL) offered by IIT Ropar and The Second Wind will receive their certificates via an online convocation ceremony on the 29th of May 2021

At the convocation ceremony the Chief Guest and Director Prof. Rajeev Ahuja, IIT Ropar said, “I congratulate the students on their success. Domain expertise in AI is a skill that will prepare you for the future, as it finds more and more applications across industry verticals. We were proud to partner with The Second Wind to deliver this course to all of you. I also take this opportunity to thank the dedicated faculty at IIT Ropar and The Second Wind for navigating the students through this complex program. We at IIT Ropar are dedicated to offer more such innovative programs as we leverage our efficient operations and advanced technology. Our ultimate goal is to transform education with innovation and technology”.

Commenting at the ceremony, Anish Srikrishna, CEO Times Professional Learning said, “With the rapid emergence of AI, this course will prepare working professionals to meet the dynamic challenges at their workplace and help enhance their career. The Second wind will continue to play a leadership role towards building competencies that meet the requirements of the industry. I heartily congratulate all the students and wish them a very successful career ahead.”

Also present on the occasion were Dr. Rohit Sharma, Program Director – PGCAIDL, Co-ordinator ITJRC, Associate Professor – Department of Electrical Engineering IIT Ropar, Mr. Varun Dhamija, VP & Business Head, Online Programs – Times Professional Learning (TPL) and Mr. Mohan S, Head Technology Programs – TPL

Post Graduate Certificate in Artificial Intelligence & Deep Learning (PGCAIDL) is a 9 months program with live instructor-led sessions by top IIT faculty and industry experts. This artificial intelligence online course is designed for working professionals with weekend classes to impart AI learning.

About IIT Ropar

IIT Ropar is an engineering, science, and technology higher education institute located in Rupnagar, Punjab, India, imparting state-of-the-art technical education in a variety of fields. Its emphasis on promoting cutting-edge research and high quality publications is the key to its recognition in the international research community. It is a well reputed institute, having secured high ranks in ‘Times Higher Education (THE) World University Rankings 2020’, ‘QS India Rankings 2020’, and Union HRD Ministry’s ‘National Institutional Ranking Framework (NIRF)’, among others.

About The Second Wind (TSW)

The Second Wind (TSW) is an initiative by Times Professional Learning for working professionals who are keen to hone their skills to add value to their work for a more rewarding career. TSW’s passion for excellence and a brief that “Executive Education Empowers” works hand-in-hand with the organisation’s aim to impart knowledge to business professionals nationwide.

New Deep Learning Discovery Paves Way for AI Interpretation of Brainwave Data

A new paper published in the Journal of Neural Engineering shows the successful first application of self-supervised learning, a very promising recent approach to train deep neural networks, to directly learn what EEG looks like, without using any labelled data.

TORONTO , Jan. 25, 2021 /PRNewswire/ -- Interpreting the results of electroencephalogram (EEG) graphs, which are used to visualize brain activity of everything from meditation to neurological disorders, is one of the greatest challenges facing brain researchers. Machine learning has the potential to relieve some of this burden, but EEG data is extremely multidimensional and can be expensive, and time-consuming to annotate. It also requires the deep expertise of neurologists and sleep experts. This means there are typically not enough labelled examples for supervised deep neural networks to learn from in order to create an efficient AI.

While labelled-EEGs identifying sleep stages and brain activity are scarce, there is ample unlabeled data that exists. In his new paper, Uncovering the Structure of Clinical EEG Signals with Self-supervised Learning , Interaxon Inc researcher Hubert Banville and researchers at Université Paris-Saclay, University of Helsinki , and Max Planck Institute applied self-supervised learning to extract features from unlabeled EEGs. Banville found that when limited numbers of labelled data were available, his self-supervised learning approach outperformed traditional supervised learning methods that rely purely on labelled data. These results were obtained on two very different EEG classification problems: identifying sleep stages in overnight recordings, and detecting EEG pathologies. Moreover, without access to any labelled data, his approach also uncovered a fascinating structure in the data that relates to clinical information such as sleep stages, pathology and age. 

Put into practice, these intelligent new approaches will allow the scientific community, including researchers at Interaxon Inc., the developer of Muse® headbands that has one of the largest brain data (EEG) collections in the world, to leverage, mine and utilize large amounts of unlabelled EEG data to efficiently discover relevant information in EEG. This has potential to improve the performance of algorithms used in everything from consumer sleep and wellness support tools like Muse S , to swifter diagnosis of neurological disorders.

Read the full paper, methodology and results here . For interviews and more information, contact Talia Noya at press@choosemuse.com.

About Interaxon Inc. (aka Muse) :

Our team of neuroscientists, meditation teachers, and engineers develop state-of-the-art experiences using research-grade EEG technology. Our goal is to help individuals build a rewarding meditation and sleep practice to  live healthier, happier, more connected lives through human-centered technology. Our award-winning neurofeedback devices, and premium content offering of guided meditations with responsive learning functionality, help users meditate and improve sleep hygiene by providing real-time audio feedback on their meditative state through the Muse® companion app. We make the intangible, tangible. More information about Muse® is available at www.choosemuse.com .

Milestones we're excited about:

• # of employees: 45+ 
• Offices: Toronto, Ontario Sessions of meditation with Muse ®: Over 165 million minutes to date and currently one of the largest brain data (EEG) collections in the world. 
• Research institutions using Muse®: The Mayo Clinic, NASA, Harvard , MIT , U of T, UCL, UCSD, Inria, UVic, UBC, and many more. 
• Meditating with Muse® works: A recent study at the Catholic University of Milan showed that four weeks with Muse® significantly reduced stress as well as potentially influenced beneficial neuroplastic changes in users' brains, compared to controls. Original published papers from the Balconi Lab study can be found here and here .

SOURCE Muse by Interaxon Inc.

IISc and TalentSprint Announce PG Level Deep Learning Specialization Program

• For aspiring and practicing AI and ML professionals keen to build Deep Learning expertise and master applications in speech, text, image and video.
• Taught through live interactive online classes by inter-disciplinary group of renowned IISc faculty, hands on learning on curated projects with industry mentor support.

The Indian Institute of Science (IISc) and TalentSprint today announced the launch of a PG level Advanced Certification Program in Deep Learning. This 10-month executive education program is best suited for aspiring and practicing Artificial Intelligence and Machine Learning professionals who want to build expertise in Deep Learning and master applications in speech, text, image and video, including hardware optimization. A team of 10+ interdisciplinary faculty from IISc and TalentSprint will teach participants through live and interactive online classes and mentor them as they learn hands-on through curated industry capstones, including the option of bringing their own projects. The program will enable participants to establish a portfolio that demonstrates their learnings. Enrolments for the first cohort are open now and classes will commence in March 2021. Please visit the program page (link) to know more and apply.

According to a report by Grand View Research, the global Deep Learning market is expected to reach USD 10.2 billion by 2025. Improvements in algorithms and advancements in Deep Learning chipsets are driving this growth. Deep Learning is penetrating various sectors at an unprecedented speed, including financial services, healthcare, software, retail, manufacturing, etc. It is being used in investment modeling, fraud detection, designing autonomous cars, virtual assistants, super-computing, customer relationship management (CRM) systems, facial recognition systems, among others. Deep Learning will continue to make an impact in both professional and personal spaces and create several job opportunities.

Announcing the program, Prof Chiranjib Bhattacharya, Professor and Chair of the Department of Computer Science & Automation, and Dean of the Advanced Deep Learning Program, said: “Deep Learning is increasingly being used to extract valuable insights from enormous amounts of data, build innovative products and improve customer experience, thereby enhancing revenue opportunities. This has led to a massive growth in need for professionals with expertise in Deep Learning. This program will fulfill that need. Our team of research faculty will teach and mentor participants and help them build expertise in both the fundamentals and applications of Deep Learning.”

Dr. Santanu Paul, Co-Founder and CEO, TalentSprint said: “At TalentSprint, our focus has always been to offer transformational high-end and deep-tech learning programs. We are happy to announce our third program with IISc, India’s leading institution, in Deep Learning. This program has been specifically designed for working professionals keen to unleash the potential of deep learning in their careers.  This pracademic program will bring together lectures by top research faculty at IISc and hands-on labs and projects supported by industry mentors.” 

The Advanced Certification Program in Deep Learning is open for applications for the first cohort, which starts in March 2021. Please visit the program page (link) to know more and apply.
 

Visa Harnesses Real-Time Deep Learning to Enhance Transaction Processing

Visa Inc. (NYSE: V) today announced Visa Smarter Stand-in Processing (Smarter STIP), a new capability that uses real-time artificial intelligence (AI) to help financial institutions manage transaction authorizations when service disruptions occur.

Using deep learning to analyze past transactions, Smarter STIP generates informed decisions to approve or decline transactions on behalf of issuers in the event that their systems go offline. Visa pioneered the use of AI and neural networks to prevent fraud—Smarter STIP builds on that track record, and is the first in a suite of new AI-powered innovations coming to VisaNet, Visa’s global processing network.

“Consumers expect instant and uninterrupted access to their funds—failure to meet that expectation can be costly,” says Jack Forestell, executive vice president and chief product officer, Visa. “Stand-in processing has always been a vital assurance to our clients that when the unexpected occurs, we’ve got their back. By adding AI, that capability becomes smarter, stronger, and more dynamic.”

Addressing the challenge of downtime

Whether it’s scheduled maintenance or an unexpected outage, service downtime is disruptive for financial institutions and their customers. When stand-in processing is not in place, the potential impacts can be profound: lost revenue due to failed transactions, poor cardholder experience resulting in call volume spikes to customer service centers, damage to reputation, and even scrutiny from regulators.

Visa’s new Smarter STIP service builds on Visa’s existing STIP capability by using deep learning to analyze past transactions down to the cardholder level. Thus the transaction decision that Smarter STIP provides is based on unique insights derived from the cardholder’s past purchasing behavior, rather than solely on static rules applied across an entire card portfolio. With this added intelligence, Visa is able to provide a transaction decision on the issuer’s behalf that more closely mirrors the issuer’s own decision making process—with the potential to decrease transaction declines for cardholders by up to 50% in some cases.¹

Harnessing AI for the Payments Ecosystem

Visa Smarter STIP is one of several new real-time deep learning capabilities designed to solve long-standing challenges and pain points for buyers, sellers, and financial institutions.

These enhanced services are made possible by the investments Visa has made to optimize its core infrastructure. This includes a scalable, high-performing, GPU-based platform designed to support the rapid deployment of deep learning capabilities.

“AI is one of the most transformative technologies of our time, giving us the power to build data-driven products that deliver new levels of insight and personalization. With one of the largest and richest data sets in the world, Visa is extremely well positioned to unlock the full potential of AI for our clients and cardholders, and at massive scale,” says Rajat Taneja, president of technology at Visa. “We’ve made significant investments in our infrastructure, making it possible to harness this data more deeply across all facets of VisaNet. Our data, combined with low-latency architecture and our unique modeling capabilities, make our technology platform among the most powerful for AI.”​

Global adoption already underway

Smarter STIP will be available in October 2020 and has generated interest from a range of financial institutions around the world.

“At UBS, we believe in building a better future for our customers by harnessing the power of innovation and collaboration—a continuous process that relies on innovative thinking and transformative technologies and capabilities, developed both in-house and by our partners,” says Stefan Uebelhart, Head of UBS Card Center. “We are very pleased to launch Visa’s Smarter STIP service, as this is an example where the power of collaboration, transformative technologies and innovation come together to provide an improved experience for our cardholders. In our quest to build a better future for our customers, we see Visa as a valued partner and intend to work together in the innovation space.”

“As a leader in interest-free shopping platforms in Australia and New Zealand, Latitude Financial is constantly exploring new technology and innovation so we can deliver the best possible service to our customers,” said Latitude’s Chief Commercial Officer, Paul Varro. “Visa’s Smarter STIP service helps us substantially improve transaction approval rates. This is a great outcome, cleverly bringing together AI with our historical authorisation learnings, new datasets, and Visa’s experience in transaction processing to provide a rich customer experience, all the time.”

About Smarter STIP

Visa created the Smarter STIP capability to address a key pain point for issuers and issuer processors: the suboptimal experience for cardholders during service disruptions. Visa’s novel approach to stand-in processing, recognized through a US Patent in 2019, improves on the existing STIP capability through real-time evaluation of cardholder-level features. The model relies on:

• Multiple recurrent neural networklayers with millions of parameters.


• Billions of historical records to train the model, resulting in an average 95% accuracy in emulating an issuer’s likely decision in Visa’s sample tests².


• Continuous learning from real time transaction and outage data, allowing the service to adapt and improve as an Issuers’ behavior changes over time​.

Visa Smarter STIP does not require any technical change for clients to participate. Issuers connected to Visa DPS will be enabled for Smarter STIP when the service launches globally in October 2020.

About Visa Inc.

Visa Inc. (NYSE: V) is the world’s leader in digital payments. Our mission is to connect the world through the most innovative, reliable and secure payment network - enabling individuals, businesses and economies to thrive. Our advanced global processing network, VisaNet, provides secure and reliable payments around the world, and is capable of handling more than 65,000 transaction messages a second. The company’s relentless focus on innovation is a catalyst for the rapid growth of digital commerce on any device for everyone, everywhere. As the world moves from analog to digital, Visa is applying our brand, products, people, network and scale to reshape the future of commerce. For more information, visit About Visa,visa.com/blog and @VisaNews.

---

¹ Based on Visa internal analysis comparing existing and Smarter STIP approval rates of all transactions in Q1 CY2020 for a single US Issuer.
² This number was measured by Visa’s internal simulation of the Smarter STIP model offline on transactions for all Visa BINs globally in Q4 2019. On average, the Smarter STIP model made the same approve/decline decision as the issuers 95% of the time, for transactions across all regions.

View source version on businesswire.com: https://www.businesswire.com/news/home/20200826005289/en/

IIT Roorkee launches Deep Learning Course at CloudxLab.com to Upskill Youth During COVID-19 Lockdown

In an endeavour to upskill the youth and promote e-learning during the COVID-19 lockdown, IIT Roorkee has launched an Advanced Certification Course on Deep Learning at Cloudxlab.com.

The launch comes in the wake of the current economic crisis which underscores the significance of technical skills to tackle the global slowdown. It follows the signing of the Memorandum of Understanding (MoU) between IIT Roorkee and Cloudxlab.com, a US-based ed-tech company, for offering a series of instructor-led as well as self-paced executive online courses.

“COVID-19 has triggered a nationwide lockdown. This is the best time for the youth as well as others to upskill themselves. This initiative will be attractive for users who aim to excel in the technical field. The partnership with CloudxLab.com will amplify our reach to offer the latest knowledge in tune with the needs of the industry,” said Prof. Ajit K. Chaturvedi, Director, IIT Roorkee.

The course is a new addition to a host of pre-existing courses such as Artificial Intelligence, Machine Learning, Data Science and many more. It will be delivered by the faculty of IIT Roorkee and other experts from the industry. The classes will be streamed online through a live video. Upon completion of the course, users will receive a certification from IIT Roorkee.

“Technology continues to advance rapidly even today, and this is perhaps the best time for professionals to expand their horizons and learn new technologies,” said Manish Shrikande, Dean, Sponsored Research and Industrial Consultancy, IIT Roorkee.

The initiative will equip students as well as professionals who wish to utilize this lockdown period in upskilling themselves.

About Cloudxlab.com

(https://cloudxlab.com/course/certificate-course-aritifical-intelligence-deep-learning-iit-roorkee/

CloudxLab, an ed-tech venture headquartered in the US, provides an online learning platform that upskills engineers in deep tech and makes them employable & future-ready.

About IIT Roorkee 

IIT Roorkee is an institute of national importance imparting higher education in engineering, sciences, management, architecture and planning, and humanities and social sciences. Since its establishment in 1847, the Institute has played a vital role in providing technical human resources and know-how to the country.

Manipulated Video & Audio Made using 'Deepfakes' Poses Threat to Elections

A video on social media shows a high-ranking U.S. legislator declaring his support for an overwhelming tax increase. You react accordingly because the video looks like him and sounds like him, so certainly it has be him.

The term "fake news" is taking a much more literal turn as new technology is making it easier to manipulate the faces and audio in videos. The videos, called deepfakes, can then be posted to any social media site with no indication they are not the real thing.

Edward Delp, director of the Video and Imaging Processing Laboratory at Purdue University, says deepfakes are a growing danger with the next presidential election fast approaching.

“It’s possible that people are going to use fake videos to make fake news and insert these into a political election,” said Delp, the Charles William Harrison Distinguished Professor of Electrical and Computer Engineering. “There’s been some evidence of that in other elections throughout the world already.

“We’ve got our election coming up in 2020 and I suspect people will use these. People believe them and that will be the problem.”

The videos pose a danger to swaying the court of public opinion through social media, as almost 70 percent of adults indicate they use Facebook, usually daily. YouTube boasts even higher numbers, with more than 90 percent of 18- to 24-year-olds using it.

Delp and doctoral student David Güera have worked for two years on video tampering as part of a larger research into media forensics. They’ve worked with sophisticated machine learning techniques based on artificial intelligence and machine learning to create an algorithm that detects deepfakes.

A YouTube video is available at https://youtu.be/aWKBWoDtR8k.

Detecting Deep Fakes Video through Media Forensics

[tnm_video layout="mnmd-post-media-wide"]https://www.youtube.com/watch?v=aWKBWoDtR8k[/tnm_video]

Late last year, Delp and his team’s algorithm won a Defense Advanced Research Projects Agency (DARPA) contest. DARPA is an agency of the U.S. Department of Defense.

“By analyzing the video, the algorithm can see whether or not the face is consistent with the rest of the information in the video,” Delp said. “If it’s inconsistent, we detect these subtle inconsistencies. It can be as small as a few pixels, it’s can be coloring inconsistencies, it can be different types of distortion.”

“Our system is data driven, so it can look for everything – it can look into anomalies like blinking, it can look for anomalies in illumination,” Güera said, adding the system will continue to get better at detecting deepfakes as they give it more examples to learn from.

The research was presented in November at the 2018 IEEE International Conference on Advanced Video and Signal Based Surveillance.

Deepfakes also can be used to fake pornography video and images, using the faces of celebrities or even children.

Delp said early deepfakes were easier to spot. The techniques couldn’t recreate eye movement well, resulting in videos of a person that didn’t blink. But advances have made the technology better and more available to people.

News organizations and social media sites have concerns about the future of deepfakes. Delp foresees both having tools like his algorithm in the future to determine what video footage is real and what is a deepfake.

“It’s an arms race,” he said. “Their technology is getting better and better, but I like to think that we’ll be able to keep up."

Top Image - Blog.avira.com

AI to Steal Fashion Model Jobs?, New AI Able to Generate Entire Bodies of People Who Don’t Exist

A new Artificial Intelligence (AI) algorithm can now generate high-resolution, photorealistic images of people from scratch which do not exist.

Called as Generative Adversarial Network (GAN), this new class of machine learning systems can generate photographs that look at least superficially authentic to human observers having many realistic characteristics -- faces, hair, outfits, and all. The only thing that these AI-generated people "Do Not Exist" in reality and fashion brands or advertising agencies can use these photogenic AI generated models without paying for lights or a catering budget, which means ditching the fashion models that are actual humans.

[video width="3072" height="1024" mp4="https://www.indianweb2.com/wp-content/uploads/2019/05/full-body-model.mp4"][/video]

Datagrid, a tech company housed on the campus of Japan’s Kyoto University, has developed an AI called "automatic whole-body model generation AI" by applying the GAN based deep learning which learn a large number of whole body model images to generate non-existent whole body model images with high resolution (1024 x 1024) and high quality, which had previously been difficult.

The new kind of AI algorithm typically used to churn out new imitations of something that exists in the real world, whether they be video game levels or images that look like hand-drawn caricatures.

[tnm_video layout="mnmd-post-media-wide"]https://www.youtube.com/watch?v=8siezzLXbNo[/tnm_video]

The Japan-based tech firm had earlier developed idle automatic generation AI in June 2018. However, idle auto generation AI generates an image only for the face area, so it did not have enough expressive power. Therefore, in order to enhance the expressive power of the generated person, Datagrid has been working on two research and development of "whole body generation" and "motion generation". The high-precision whole-body generation model has no precedent, and was a challenging research and development. The company however has now succeeded in stably generating high-resolution (1024 × 1024) whole-body model images.

In its Future Prospects, the company said in a press release that it will further improve the accuracy of the whole-body model automatic generation AI and research and develop the motion generation AI. Additionally, Datagrid will conduct demonstration experiments with advertising and apparel companies to develop functions required for actual operation.

Ethical Disadvantages

According to critics, GANs potential in human image synthesis can be used for sinister purposes such as to produce fake and/or incriminating photographs and videos and similar algorithms could be misused to undermine public trust in digital media.

A similar AI technology called "Deepfakes", which can doctored images and videos that can be used to generate propaganda and deceptive media like fictionalized political speeches or pornography, is difficult to counter so much so that Canadian Broadcasting Corporation has called it as a matter of national security.

Nasscom To Develop 2000 Member World-Class Faculty in Artificial Intelligence, Machine Learning and Deep Learning

The National Association of Software and Services Companies (NASSCOM), today launched the ‘AI Foundation for Faculty Development’ program through its Centre of Excellence for Data Science & Artificial Intelligence (CoE DS&AI) and FutureSkills platform. Broadening its resolve to democratize AI learning, the Faculty Development Program offers a comprehensive and a well curated schedule. It aims to provide the faculty in AI with the necessary exposure to real-life business problems by working closely with industry experts and providing access to latest tools and technologies in the AI space.

The program will be delivered as an integrated platform consisting of 125 hours of learning including approximately 100 hours of self-paced online learning in areas of Predictive Modelling, Programming Languages, Time Series Forecasting, Machine Learning, Optimization Techniques, Visualization, Natural Language Processing and Image & Video Analytics. This online component will be delivered through the NASSCOM FutureSkills platform which brings together the best-in-class global learning and content providers to offer learners free and deep learning content in emerging technologies such as AI. Upon successful completion of the online learning, the participating faculty will be required to undergo 25 hours of hands-on, lab-based training by industry experts and academic partners on cutting-edge business problems. Post the hands-on lab based training, the participants will be certified by NASSCOM.

The objective of this initiative is to develop an eco-system of world-class faculty in the AI Space, which translates into world-class talent who are ready to be absorbed by the industry in to their workforce. The program will be launched initially for technical colleges and universities of Karnataka and will expand across the country eventually. As part of this program, NASSCOM CoE – DSAI aims to upskill nearly 2000 faculty members in Artificial Intelligence, Machine Learning and Deep Learning through multiple interventions and impact nearly 25,000 undergraduates pursuing AI related courses.

Expressing his thoughts on the initiative, Gaurav Gupta, Principal Secretary for Industries, IT, BT & ST, Government of Karnataka, said “In order to catalyze innovation and excellence in data science and artificial intelligence, there is an urgent need to invest in talent development. Globally, there is a shortage of AI talent. Such investments in enhancing the capacities of faculty and young talent will enable Karnataka to become a global powerhouse in AI. The Faculty Development Program in AI will pave the way for a nationwide development of faculty in AI, ML, and advanced deep learning impacting the future talent the industry seeks to recruit and provide an equal opportunity to all”.

Sharing his delight on the launch of the program, Amit Aggarwal, CEO Future Skills, NASSCOM said “As adoption of Automation and Artificial Intelligence builds momentum across the world, millions of people worldwide will have to upgrade their skills. The skills that were once relevant are no longer going to be effective enough for a dynamic career. Addressing the skills gap requires a calibrated approach and re-skilling of faculty is one of the critical drivers for this change. Through this programme, we hope to co-create a faculty development ecosystem that strengthens our ability to prepare our students for the jobs of future.”

Mohanakrishnan, Head – CoE DSAI said, AI is fast transforming our lives, industries and society. The Faculty Development Programs aims to catalyse knowledge of faculty in AI with technology insights and a practical, in-depth understanding of real business problems & application across industries.”

In the recent times, artificial intelligence has emerged from the confines of scientific labs and found various uses to impact our everyday life. It is expected that the data science and artificial intelligence industry in India (IT and non-IT industries) will be worth US $16 billion by 2025. As per a recent study by NASSCOM, there is a likely shortage of 2,38,000 data science & artificial intelligence professionals over the next three years in the country.

The above news was first published at Express Computer

Japan's E-Commerce Giant Opens R&D Centre in India for Research in Deep Learning, Computer Vision

Tokyo, Japan-based e-commerce and internet giant Rakuten, Inc., has opened its India research and development (R&D) institute here to focus on deep learning, computer vision and robotics for warehouse logistics applications.

At its operations centre in Bengaluru, the company has opened Rakuten Institute of Technology (RIT) as its research centre in India that will combine creativity and innovation with emerging technologies to fuel the growth of e-commerce. The centre is Rakuten's sixth R&D centre across the globe as its five other research centres are in Tokyo, Paris, Singapore, Boston and San Mateo in the US.

AT RIT Bangalore, the applied research on deep learning will be to detect and predict financial fraud, developing applications in computer vision for e-commerce and medical imagining and advancement of automation to improve efficiency in warehouse robotics and logistics.

The Bangalore R&D facility will also be a hub for collaboration with Indian universities to provide researchers opportunities to study internet technologies and develop services for the Rakuten group companies the world over.

The company has appointed Sunil Gopinath as the Institute's first Chief Executive. Gopinath was earlier working at Flipkart and eBay.

"India has a large technical talent and its researchers have an understanding of combining academic research with business domain applications to deliver pioneering business solutions," said Gopinath.

The RIT also partners with Hidehiko Masuhara of the Tokyo Institute of Technology and Ruby computer programming language creator Yukihiro Masumoto serves as a fellow.

The RIT's first centre was set up in Japan in 2006 to keep pace with the evolution of the cyber world and the rapid growth of internet services globally.

Masaya Mori, who is global head at RIT, said, "India has a vibrant computer science research community in Bengaluru and we look forward to collaborate with academic researchers to discover and develop new technology applications."

Rakuten is looking to move the centre of gravity of its technology operations to India and seeking to grab a share of Indian e-commerce industry that has been growing at an average annual rate of 34% since 2009.

Notably, Flipkart's parenr Walmart on other hand has had also partnered with Rakuten, earlier this year. As part of the deal, Walmart will start selling Aura-branded e-readers made by Kobo, a division of Rakuten, in stores and online at Walmart.com. Kobo is one of the Amazon Kindle's competitors.

Source - Yahoo Finance

[Top Image - New Straits Times]

Google Building An AI Tool To Forecast When A Patient Will Die

A Google team led by Alvin Rajkomar and Eyal Oren in the company's headquarter Mountain View, California, US, has developed a data processing pipeline for transforming electronic health record (EHR) files into a standardized format. They then applied deep learning models to data from 216,221 adult patients hospitalized for at least 24 hour each at two academic medical centers, and showed that their algorithm could accurately predict risk of mortality, hospital readmission, prolonged hospital stay and discharge diagnosis.

With this AI driven tool, Google can forecast a host of patient outcomes, including how long people may stay in hospitals, their odds of re-admission and chances they will soon die.

In one of a case study as proof-of-concept, Google told a real-life incidence as:

A woman with late-stage breast cancer came to a city hospital, fluids already flooding her lungs. She saw two doctors and got a radiology scan. The hospital’s computers read her vital signs and estimated a 9.3 percent chance she would die during her stay.

Then came Google’s turn. An new type of algorithm created by the company read up on the woman -- 175,639 data points -- and rendered its assessment of her death risk: 19.9 percent. She passed away in a matter of days.

Medical experts too are impressed by Google's ability to throughly examine the data, which otherwise was out of reach lying stagnat in PDFs or scribbled on old charts. The neural network of Google gobbled up all this unruly information & records and then throw out predictions. And it did it far faster and more accurately than existing techniques. Google’s system even showed which records led it to conclusions.

Google's health research unit working on this tool -- being referred to as Medical Brain -- is now working on a slew of AI tools that can predict symptoms and disease.

Also Read - Microsoft Aims To Use Drone Technology To Catch Mosquitoes & Predict Diseases Outbreaks

As much as 80 percent of the time spent on today’s predictive models goes to the “scut work” of making the data presentable, said Nigam Shah, an associate professor at Stanford University, who co-authored Google’s research paper, published in the journal Nature. Google’s approach avoids this. "You can throw in the kitchen sink and not have to worry about it,” Shah said.

Inside the company, there’s a lot of excitement about the initiative. "They’ve finally found a new application for AI that has commercial promise," one Googler says. Since Alphabet Inc.’s Google declared itself an “AI-first” company in 2016, much of its work in this area has gone to improve existing internet services. The advances coming from the Medical Brain team give Google the chance to break into a brand new market -- something co-founders Larry Page and Sergey Brin have tried over and over again.

Software in health care is largely coded by hand these days. In contrast, Google’s approach, where machines learn to parse data on their own, “can just leapfrog everything else,” said Vik Bajaj, a former executive at Verily, an Alphabet health-care arm, and managing director of investment firm Foresite Capital. “They understand what problems are worth solving," he said. "They’ve now done enough small experiments to know exactly what the fruitful directions are.”

Dean envisions the AI system steering doctors toward certain medications and diagnoses. Another Google researcher said existing models miss obvious medical events, including whether a patient had prior surgery. The person described existing hand-coded models as “an obvious, gigantic roadblock” in health care. The person asked not to be identified discussing work in progress.

For all the optimism over Google’s potential, harnessing AI to improve health-care outcomes remains a huge challenge. Other companies, notably IBM’s Watson unit, have tried to apply AI to medicine but have had limited success saving money and integrating the technology into reimbursement systems.

Google has long sought access to digital medical records, also with mixed results. For its recent research, the internet giant cut deals with the University of California, San Francisco, and the University of Chicago for 46 billion pieces of anonymous patient data. Google’s AI system created predictive models for each hospital, not one that parses data across the two, a harder problem. A solution for all hospitals would be even more challenging. Google is working to secure new partners for access to more records.

A deeper dive into health would only add to the vast amounts of information Google already has on us. "Companies like Google and other tech giants are going to have a unique, almost monopolistic, ability to capitalize on all the data we generate," said Andrew Burt, chief privacy officer for data company Immuta. He and pediatric oncologist Samuel Volchenboum wrote a recent column arguing governments should prevent this data from becoming "the province of only a few companies," like in online advertising where Google reigns.

Google is treading carefully when it comes to patient information, particularly as public scrutiny over data-collection rises. Last year, British regulators slapped DeepMind, another Alphabet AI lab, for testing an app that analyzed public medical records without telling patients that their information would be used like this. With the latest study, Google and its hospital partners insist their data is anonymous, secure and used with patient permission. Volchenboum said the company may have a more difficult time maintaining that data rigor if it expands to smaller hospitals and health-care networks.

Still, Volchenboum believes these algorithms could save lives and money. He hopes health records will be mixed with a sea of other stats. Eventually, AI models could include information on local weather and traffic -- other factors that influence patient outcomes. "It’s almost like the hospital is an organism," he said.

Few companies are better poised to analyze this organism than Google. The company and its Alphabet cousin, Verily, are developing devices to track far more biological signals. Even if consumers don’t take up wearable health trackers en masse, Google has plenty of other data wells to tap. It knows the weather and traffic. Google’s Android phones track things like how people walk, valuable information for measuring mental decline and some other ailments. All that could be thrown into the medical algorithmic soup.

Medical records are just part of Google’s AI health-care plans. Its Medical Brain has unfurled AI systems for radiology, ophthalmology and cardiology. They’re flirting with dermatology, too. Staff created an app for spotting malignant skin lesions; a product manager walks around the office with 15 fake tattoos on her arms to test it.

Dean, the AI boss, stresses this experimentation relies on serious medical counsel, not just curious software coders. Google is starting a new trial in India that uses its AI software to screen images of eyes for early signs of a condition called diabetic retinopathy. Before releasing it, Google had three retinal specialists furiously debate the early research results, Dean said.

Over time, Google could license these systems to clinics, or sell them through the company’s cloud-computing division as a sort of diagnostics-as-a-service. Microsoft Corp., a top cloud rival, is also working on predictive AI services. To commercialize an offering, Google would first need to get its hands on more records, which tend to vary widely across health providers. Google could buy them, but that may not sit as well with regulators or consumers. The deals with UCSF and the University of Chicago aren’t commercial.

For now, the company says it’s too early to settle on a business model. At Google’s annual developer conference in May, Lily Peng, a member of Medical Brain, walked through the team’s research outmatching humans in spotting heart disease risk. "Again," she said. "I want to emphasize that this is really early on."

Source - Bloomberg

[Top Image - db.dudyadityawan.com]

Google Building An AI Tool To Forecast When A Patient Will Die

Representative

A Google team led by Alvin Rajkomar and Eyal Oren in the company's headquarter Mountain View, California, US, has developed a data processing pipeline for transforming electronic health record (EHR) files into a standardized format. They then applied deep learning models to data from 216,221 adult patients hospitalized for at least 24 hour each at two academic medical centers, and showed that their algorithm could accurately predict risk of mortality, hospital readmission, prolonged hospital stay and discharge diagnosis.

With this AI driven tool, Google can forecast a host of patient outcomes, including how long people may stay in hospitals, their odds of re-admission and chances they will soon die.

In one of a case study as proof-of-concept, Google told a real-life incidence as:

A woman with late-stage breast cancer came to a city hospital, fluids already flooding her lungs. She saw two doctors and got a radiology scan. The hospital’s computers read her vital signs and estimated a 9.3 percent chance she would die during her stay.

Then came Google’s turn. An new type of algorithm created by the company read up on the woman -- 175,639 data points -- and rendered its assessment of her death risk: 19.9 percent. She passed away in a matter of days.

Medical experts too are impressed by Google's ability to throughly examine the data, which otherwise was out of reach lying stagnat in PDFs or scribbled on old charts. The neural network of Google gobbled up all this unruly information & records and then throw out predictions. And it did it far faster and more accurately than existing techniques. Google’s system even showed which records led it to conclusions.

Google's health research unit working on this tool -- being referred to as Medical Brain -- is now working on a slew of AI tools that can predict symptoms and disease.

Also Read - Microsoft Aims To Use Drone Technology To Catch Mosquitoes & Predict Diseases Outbreaks

As much as 80 percent of the time spent on today’s predictive models goes to the “scut work” of making the data presentable, said Nigam Shah, an associate professor at Stanford University, who co-authored Google’s research paper, published in the journal Nature. Google’s approach avoids this. "You can throw in the kitchen sink and not have to worry about it,” Shah said.

Inside the company, there’s a lot of excitement about the initiative. "They’ve finally found a new application for AI that has commercial promise," one Googler says. Since Alphabet Inc.’s Google declared itself an “AI-first” company in 2016, much of its work in this area has gone to improve existing internet services. The advances coming from the Medical Brain team give Google the chance to break into a brand new market -- something co-founders Larry Page and Sergey Brin have tried over and over again.

Software in health care is largely coded by hand these days. In contrast, Google’s approach, where machines learn to parse data on their own, “can just leapfrog everything else,” said Vik Bajaj, a former executive at Verily, an Alphabet health-care arm, and managing director of investment firm Foresite Capital. “They understand what problems are worth solving," he said. "They’ve now done enough small experiments to know exactly what the fruitful directions are.”

Dean envisions the AI system steering doctors toward certain medications and diagnoses. Another Google researcher said existing models miss obvious medical events, including whether a patient had prior surgery. The person described existing hand-coded models as “an obvious, gigantic roadblock” in health care. The person asked not to be identified discussing work in progress.

For all the optimism over Google’s potential, harnessing AI to improve health-care outcomes remains a huge challenge. Other companies, notably IBM’s Watson unit, have tried to apply AI to medicine but have had limited success saving money and integrating the technology into reimbursement systems.

Google has long sought access to digital medical records, also with mixed results. For its recent research, the internet giant cut deals with the University of California, San Francisco, and the University of Chicago for 46 billion pieces of anonymous patient data. Google’s AI system created predictive models for each hospital, not one that parses data across the two, a harder problem. A solution for all hospitals would be even more challenging. Google is working to secure new partners for access to more records.

A deeper dive into health would only add to the vast amounts of information Google already has on us. "Companies like Google and other tech giants are going to have a unique, almost monopolistic, ability to capitalize on all the data we generate," said Andrew Burt, chief privacy officer for data company Immuta. He and pediatric oncologist Samuel Volchenboum wrote a recent column arguing governments should prevent this data from becoming "the province of only a few companies," like in online advertising where Google reigns.

Google is treading carefully when it comes to patient information, particularly as public scrutiny over data-collection rises. Last year, British regulators slapped DeepMind, another Alphabet AI lab, for testing an app that analyzed public medical records without telling patients that their information would be used like this. With the latest study, Google and its hospital partners insist their data is anonymous, secure and used with patient permission. Volchenboum said the company may have a more difficult time maintaining that data rigor if it expands to smaller hospitals and health-care networks.

Still, Volchenboum believes these algorithms could save lives and money. He hopes health records will be mixed with a sea of other stats. Eventually, AI models could include information on local weather and traffic -- other factors that influence patient outcomes. "It’s almost like the hospital is an organism," he said.

Few companies are better poised to analyze this organism than Google. The company and its Alphabet cousin, Verily, are developing devices to track far more biological signals. Even if consumers don’t take up wearable health trackers en masse, Google has plenty of other data wells to tap. It knows the weather and traffic. Google’s Android phones track things like how people walk, valuable information for measuring mental decline and some other ailments. All that could be thrown into the medical algorithmic soup.

Medical records are just part of Google’s AI health-care plans. Its Medical Brain has unfurled AI systems for radiology, ophthalmology and cardiology. They’re flirting with dermatology, too. Staff created an app for spotting malignant skin lesions; a product manager walks around the office with 15 fake tattoos on her arms to test it.

Dean, the AI boss, stresses this experimentation relies on serious medical counsel, not just curious software coders. Google is starting a new trial in India that uses its AI software to screen images of eyes for early signs of a condition called diabetic retinopathy. Before releasing it, Google had three retinal specialists furiously debate the early research results, Dean said.

Over time, Google could license these systems to clinics, or sell them through the company’s cloud-computing division as a sort of diagnostics-as-a-service. Microsoft Corp., a top cloud rival, is also working on predictive AI services. To commercialize an offering, Google would first need to get its hands on more records, which tend to vary widely across health providers. Google could buy them, but that may not sit as well with regulators or consumers. The deals with UCSF and the University of Chicago aren’t commercial.

For now, the company says it’s too early to settle on a business model. At Google’s annual developer conference in May, Lily Peng, a member of Medical Brain, walked through the team’s research outmatching humans in spotting heart disease risk. "Again," she said. "I want to emphasize that this is really early on."

Source - Bloomberg

[Top Image - db.dudyadityawan.com]

Deep Learning Market To Reach $18.16 Billion By 2023

Deep Learning Market is forecast to reach $18.16 billion by 2023 from $3.18 billion in 2018 at a CAGR of 41.7% during (2018-2023) driven by improving computing power, declining hardware cost, the increasing adoption of cloud-based technology, usage in big data analytics and growing AI adoption in customer-centric services, according a report by MarketsandMarkets. Deep learning is a variety of machine learning (ML).

By vertical, deep learning in manufacturing is expected to show the most growth through 2023. Deep learning technology is used in industrial robots, machine vision systems and more to improve the process and product quality, minimize cycle time, and increase efficiency of the manufacturing process as a whole.

Rapid growth also is expected for deep-learning services. Deep-learning systems require installation, training and support, and maintenance services. Installation services allow the software to be integrated with the analytics side to help with data retrieval and to generate the desired result through computation. The use of computer systems for deep learning/AI further increases the amount of work involved in installation.

The companies that are profiled in the deep learning market report are NVIDIA (US), Intel (US), Xilinx (US), Samsung Electronics (South Korea), Micron Technology (US), Qualcomm (US), IBM (US), Google (US), Microsoft (US), and AWS (US). Some of the key start-ups included in this report are Graph core (UK), Mythic (US), Adapteva (US), and Koniku (US).

The deep learning market in APAC expected to grow at highest CAGR. This report covers the deep learning market in North America, Europe, APAC, and RoW. Rise in the adoption of deep learning technology in APAC could be attributed to the increasing applications of deep learning in media & advertising, finance, and retail sectors, among others, in technologically advancing countries such as India, China, and Japan. Growing e-commerce, online streaming, and increasing internet penetration have resulted in the growth of marketing industries. In the security vertical, with increasing incidents of cyber-attacks and a growing cyber-war in the region, organizations and governments are focusing on robust defense infrastructure.

North America accounts for a substantial share of the deep-learning market; however, the lack of technical expertise, absence of standards and protocols, and increasing complexity in hardware due to the complex algorithm used in deep-learning technology are restraining the growth of the market, MarketsandMarkets said.

Related Reading - 10 Alarming Predictions for Deep Learning in 2018

The deep learning market for manufacturing industry will witness highest growth between 2018 and 2023. Deep learning technology is used in industrial robots, machine vision systems, and others to improve the process and product quality, minimize cycle time, and increase the efficiency of the manufacturing process as a whole. North America accounts for a substantial share of the deep learning market, with the US being the major contributor. However, the lack of technical expertise and absence of standards and protocols, and increasing complexity in hardware due to complex algorithm used in deep learning technology are restraining the growth of the deep learning market.

The Deep Learning Market for services to grow at highest CAGR from 2018 to 2023. Deep learning technology is highly complex in nature requiring the implementation of sophisticated algorithms. Deep learning systems require installation; training; and support and maintenance services. Installation services allow the software to be integrated with the analytics side to enable data retrieval and generate desired result through computation. The use of computer systems for DL/AI further increases the amount of work involved in installation.

In terms of hardware, processor held the largest size of the deep learning market in 2017. Companies in industries such as healthcare and finance are investing in machine learning infrastructure. High parallel processing capabilities and improved computing power have resulted in the high adoption of GPUs in various DL applications.

The above press release was first published in Channel Partners.

According to market research firm Gartner, Deep learning will be a critical driver for demand, fraud and failure predictions by 2019. ML through deep learning is increasingly being used in predicting demand, determining deficiencies around service and product quality, detecting new types of fraud, streaming analytics on data in motion and providing predictive or even prescriptive maintenance.

Earlier this year, Chinese firm Alibaba’s iDST (Institute of Data Science of Technologies), Alibaba Group’s research arm focused on artificial intelligence, developed a deep-learning model that scored higher than a human being on a Stanford University reading-comprehension test. Notably, that was the first time that a machine has outperformed humans on such a test.

Deep Learning Market To Reach $18.16 Billion By 2023

Deep Learning Market is forecast to reach $18.16 billion by 2023 from $3.18 billion in 2018 at a CAGR of 41.7% during (2018-2023) driven by improving computing power, declining hardware cost, the increasing adoption of cloud-based technology, usage in big data analytics and growing AI adoption in customer-centric services, according a report by MarketsandMarkets. Deep learning is a variety of machine learning (ML).

By vertical, deep learning in manufacturing is expected to show the most growth through 2023. Deep learning technology is used in industrial robots, machine vision systems and more to improve the process and product quality, minimize cycle time, and increase efficiency of the manufacturing process as a whole.

Rapid growth also is expected for deep-learning services. Deep-learning systems require installation, training and support, and maintenance services. Installation services allow the software to be integrated with the analytics side to help with data retrieval and to generate the desired result through computation. The use of computer systems for deep learning/AI further increases the amount of work involved in installation.

The companies that are profiled in the deep learning market report are NVIDIA (US), Intel (US), Xilinx (US), Samsung Electronics (South Korea), Micron Technology (US), Qualcomm (US), IBM (US), Google (US), Microsoft (US), and AWS (US). Some of the key start-ups included in this report are Graph core (UK), Mythic (US), Adapteva (US), and Koniku (US).

The deep learning market in APAC expected to grow at highest CAGR. This report covers the deep learning market in North America, Europe, APAC, and RoW. Rise in the adoption of deep learning technology in APAC could be attributed to the increasing applications of deep learning in media & advertising, finance, and retail sectors, among others, in technologically advancing countries such as India, China, and Japan. Growing e-commerce, online streaming, and increasing internet penetration have resulted in the growth of marketing industries. In the security vertical, with increasing incidents of cyber-attacks and a growing cyber-war in the region, organizations and governments are focusing on robust defense infrastructure.

North America accounts for a substantial share of the deep-learning market; however, the lack of technical expertise, absence of standards and protocols, and increasing complexity in hardware due to the complex algorithm used in deep-learning technology are restraining the growth of the market, MarketsandMarkets said.

Related Reading - 10 Alarming Predictions for Deep Learning in 2018

The deep learning market for manufacturing industry will witness highest growth between 2018 and 2023. Deep learning technology is used in industrial robots, machine vision systems, and others to improve the process and product quality, minimize cycle time, and increase the efficiency of the manufacturing process as a whole. North America accounts for a substantial share of the deep learning market, with the US being the major contributor. However, the lack of technical expertise and absence of standards and protocols, and increasing complexity in hardware due to complex algorithm used in deep learning technology are restraining the growth of the deep learning market.

The Deep Learning Market for services to grow at highest CAGR from 2018 to 2023. Deep learning technology is highly complex in nature requiring the implementation of sophisticated algorithms. Deep learning systems require installation; training; and support and maintenance services. Installation services allow the software to be integrated with the analytics side to enable data retrieval and generate desired result through computation. The use of computer systems for DL/AI further increases the amount of work involved in installation.

In terms of hardware, processor held the largest size of the deep learning market in 2017. Companies in industries such as healthcare and finance are investing in machine learning infrastructure. High parallel processing capabilities and improved computing power have resulted in the high adoption of GPUs in various DL applications.

The above press release was first published in Channel Partners.

According to market research firm Gartner, Deep learning will be a critical driver for demand, fraud and failure predictions by 2019. ML through deep learning is increasingly being used in predicting demand, determining deficiencies around service and product quality, detecting new types of fraud, streaming analytics on data in motion and providing predictive or even prescriptive maintenance.

Earlier this year, Chinese firm Alibaba’s iDST (Institute of Data Science of Technologies), Alibaba Group’s research arm focused on artificial intelligence, developed a deep-learning model that scored higher than a human being on a Stanford University reading-comprehension test. Notably, that was the first time that a machine has outperformed humans on such a test.

Alibaba Group’s AI-Focussed Research Arm Develops A Deep-Learning Model That Scored Higher Than A Human Being

Alibaba’s iDST (Institute of Data Science of Technologies), Alibaba Group’s research arm focused on artificial intelligence, has developed a deep-learning model that scored higher than a human being on a Stanford University reading-comprehension test. This is the first time that a machine has outperformed humans on such a test.

The Stanford Question Answering Dataset (SQuAD) is a large-scale reading-comprehension dataset, consisting of over 100,000 question-answer pairs based on more than 500 Wikipedia articles. Participating teams are required to build machine-learning models that can provide answers to the questions in the dataset. It is perceived as the world’s top machine reading-comprehension test and attracts universities, companies and institutes ranging from Google, Facebook, IBM and Microsoft to Carnegie Mellon University, Stanford University and the Allen Research Institute.

On January 11th, the deep neural network model developed by Alibaba generated a score of 82.44 in Exact Match – providing exact answers to questions - beating the score by humans, 82.304.

The model, which leverages the innovative Hierarchical Attention Network that reads from paragraphs to sentences to words in order to locate the precise phases with potential answers, is believed to have significant commercial value. Previously, the underlying technology has been applied in Alibaba’s 11.11 Global Shopping Festival over, with machines answering a large amount of inbound inquiries during the mega-sale.

Luo Si, Chief Scientist of Natural Language Processing (NLP) at Alibaba iDST commented:

“It is our great honor to witness the milestone where machines surpass humans in reading comprehension. That means objective questions such as ‘what causes rain’ can now be answered with high accuracy by machines. We are especially excited because we believe the technology underneath can be gradually applied to numerous applications such as customer service, museum tutorials and online responses to medical inquiries from patients, decreasing the need for human input in an unprecedented way.”

“We are thrilled to see NLP research has achieved significant progress over the year. We look forward to sharing our model-building methodology with the wider community and exporting the technology to our clients in the near future,” Si added.

Alibaba’s iDST NLP team has received the best scores in previous global evaluations including the ACM CIKM cup of personalized e-commerce search, Chinese Grammar Error Diagnosis and English named entity classification task in the Text Analysis Conference.

Alibaba is committed to cutting-edge technology development. In October last year, Alibaba launched an innovative global research program, the Alibaba DAMO Academy, which will focus on topics such as machine learning, network security, visual computing and NLP, among others.

For more information about the reading comprehension test, please refer to: https://rajpurkar.github.io/SQuAD-explorer/

10 Alarming Predictions for Deep Learning in 2018

The below article was first published at Medium by Carlos E. Perez, who is an author of Artificial Intuition and the Deep Learning Playbook — Intuition Machine Inc.

The incredible breakthroughs we saw in 2017 for Deep Learning is going to carry over in a very powerful way in 2018. A lot of work coming from research will be migrating itself into everyday software applications.

Here are Carlo's predictions for 2018.

1. A Majority of Deep Learning Hardware Startups will Fail

Many Deep Learning hardware startup ventures will begin to finally deliver their silicon in 2018. These will all be mostly busts because they will forget to deliver good software to support their new solutions. These firms have as their DNA, hardware. Unfortunately, in the DL space, software is just as important. Most of these startups don’t understand software and don’t understand the cost of developing software. These firms may deliver silicon, but nothing will ever run on them!

The low hanging fruit that employs systolic array solutions has already been taken, so we won’t have the massive 10x performance upgrade that we found in 2017. Researchers will start using these tensor cores not only for inference, but also to speed up training.

Intel’s solution will continue to be delayed and will likely disappoint. The record shows that Intel was unable to deliver on a mid-2017 release and its anybody’s guess when they will ever deliver. It’s late and it’s going to be a dud.

Google will continue to surprise the world with its TPU developments. Perhaps Google gets into the hardware business by licensing their IP to other semiconductor vendors. This will make sense if they continue to be the only other real player in town other than Nvidia.

2. Meta-Learning will be the new SGD

A lot of strong research in Meta-learning appeared in 2017. As the research community collectively understands meta-learning much better, the old paradigm of Stochastic Gradient Descent (SGD) will fall in the wayside in favor of a more effective approach that combines both exploitive and exploratory search methods.

Progress in unsupervised learning will be incrementally, but it will be primarily be driven by Meta-learning algorithms.

3. Generative Models drives a New Kind of Modeling

Generative models are going to find themselves in more scientific endeavors. At present, most research is performed in generating images and speech. However we shall see this methods being incorporated in tools for modeling complex systems. One of the areas where you will see more activity in is the application of Deep Learning to economic modeling.

4. Self-Play is Automated Knowledge Creation

AlphaGo Zero and AlphaZero learn from scratch self play is a quantum leap. In my opinion, it is as the same level of impact as the discovery of Deep Learning. Deep Learning discovered universal function approximators. RL self-play discovered universal knowledge creation.

Do expect to see a lot more advances related to self-play.

5. Intuition Machines will Bridge the Semantic Gap

This is my most ambitious prediction. The semantic gap between intuition machines and rational machines will be bridged (if it is not already been bridged). Dual Process Theory ( the idea of two cognitive machinery, one that is model-free and the other that is model based) will be the more prevalent conceptualization of how new AI should be built. The notion of Artificial Intuition will be less of a fringe concept and be a more commonly accepted idea in 2018.

6. Explainability is Unachievable. We will just have to Fake It

There are two problems with explainability. The more commonly known problem is that the explanations have too many rules that a human cannot possibly grasp. The second problem, which is less known, is that there are concepts that a machine will create that will be completely alien and defy explanation. We already see this in the strategies of AlphaGo Zero and Alpha Zero. Human will observe that a move is unconventional, however they simply may not have the capacity to understand the logic behind the move.

This in my opinion is an unsolvable problem. What will happen instead is that machines are going the become very good at ‘faking explanations’. In short, the objective of explainable machines is to understand the kinds of explanations that a human can be comfortable with or will understand in an “intuitive” level. However, a complete explanation will in a majority of cases be completely inaccessible to humans.

Progress in explainability in Deep Learning will be made by creating ‘fake explanations’.

7. Deep Learning Research Information Deluge

2017 was already difficult for people following Deep Learning research. The number of submissions in the ICLR 2018 conference was around 4,000 papers. A researcher would have to cover 10 papers a day just to catch up with just this conference.

The problem is worsened in this space because there the theoretical frameworks are all works in progress. To make progress in the theoretical space, we need to seek out more advanced mathematics that can give us better insight. This is going to be a slog simply because most Deep Learning researchers don’t have the right mathematical background to understand the complexity of these kind of systems. Deep Learning needs researchers coming from complexity theory, however there are very few of these kinds of researchers.

As a consequence of too many papers and poor theory, we are simply left with the undesirable state of alchemy that we find ourselves today.

What is also missing is a general roadmap for AGI. The theory is weak, therefore the best we can do is create a roadmap with milestones that relate to human cognition. We only have framework that originate from speculative theories coming from cognitive psychology. This is a bad situation since the empirical evidence coming from these fields are spotty at best.

Deep Learning research papers will perhaps triple or quadruple in 2018.

8. Industrialization via Teaching Environments

The road to more predictable and controlled development of Deep Learning systems is through the development of Embodied Teaching environments. I discuss this in a little bit more detail here and here. If you want to find the crudest form of teaching technique, then one only has to look at how Deep Learning networks are trained. We are all due a lot more progress in this area.

Expect to see more companies revealing their internal infrastructure as to how the deploy Deep Learning at scale.

9. Conversational Cognition

The way we measure progress towards AGI is antiquated, rather a new kind of paradigm that addresses the dynamic (i.e. non-stationary) complexity of the real world is demanded. We shall see more coverage of this new area in the coming year. I will be speaking about this new Conversational Cognition paradigm in Amsterdam ( March 1–2, Information Energy 2018).

10. Ethical Use of Artificial Intelligence

The demands for more ethical use of Artificial Intelligence will increase. The population is now becoming more aware of the disastrous effects of unintended consequences of automation run amok. Simplistic automation that we today find in Facebook, Twitter, Google, Amazon etc. can lead to unwanted effects on society.

We need to understand the ethics of deploying machines that are able to predict human behavior. Facial recognition is one of the more dangerous capabilities that we have at our disposal. Algorithms that can generate media that is indistinguishable from reality is going to become a major problem. We as a society need to make the rapid transition to begin demanding that AI be used solely for the benefit of society as a whole and not as a weapon to increase inequality.

Expect to see more conversation about ethics in the coming year. However, don’t expect any new regulations. Policy makers are still years behind the curve in understanding the impact of A.I. to society. I don’t expect them to stop playing politics and start addressing the real problems of society. The U.S. population have been victims of numerous security breaches, yet there is zero new legislation or initiatives to address this serious problem. So don’t hold your breath that our dear leaders will suddenly discover new found wisdom.

India's IIS and NIT Develops AI To Identify Protesters With Their Faces Partly Covered With Scarves or Hat

If you're planning on becoming a part of a protest or a rally but don't want to reveal your identity at the same time, you might want to think about your participation again as the latter might no longer be possible. Researchers, from Cambridge University, India's National Institute of Technology, and the Indian Institute of Science have successfully developed a deep-learning algorithm that is capable of identifying an individual even when part of their face is obscured or covered by sunglasses or bandanas, as is seen during many protests, rallies and agitations.

The researchers, who are all set to present their research paper at the IEEE International Conference on Computer Vision Workshops (ICCVW) scheduled to be held from 22 Oct to 29 Oct 2017, in Venice, Italy, claim that their algorithm can correctly ID a person whose face is concealed by a scarf 67 percent of the time when they were photographed against a "complex" background, which resembles natural conditions in the real world.

The deep-learning algorithm functions in a unique way. In order to develop the system, the researchers first started out by outlining 14 key areas of the human face, a step which was then followed by training a deep-learning model to identify the 14 outlined areas. The algorithm then gets to work by connecting the points into what might seem like a "star-net structure." It then makes use of the angles between the points to identify a face. The algorithm is capable of still identifying those angles even when part of a person's face is covered, by disguises glasses, scarves and caps etc.

While the system can be of great help to authorities in identifying the miscreants and criminals whose main objective is to spread violence and unrest, but at the same time, it can also act as a hindrance to people who just want to peacefully stage a protest against a particular problem or issue but don't want to get identified.

Sharing his views on the topic, Amarjot Singh, a Ph.D student at Cambridge University and one of the researchers behind the paper said that when he was working on this method, his focus was solely on criminals who take advantage of disguises to cause unrest in the society.

However, while the researchers might not have thought about the other way the technology can be used by authoritarian regimes while they were working on the system, but it is a very important world issue because if the system falls in the wrong hands, it is capable of causing a lot of damage.

Zeynep Tufekci, a professor at the University of North Carolina and a writer at The New York Times, took to Twitter to voice her concerns about the system described in the paper. She wrote, too many worry about what AI—as if some independent entity—will do to us. Too few people worry what *power* will do *with* AI."

https://twitter.com/zeynep/status/904707522958852097

In a follow-up tweet, she drew attention towards the bigger picture that the research can result in. She writes, "This is a minor paper; narrow, conditional results. But it's the direction & this will be done with nation-state data—not by grad students."

While the algorithm described in the paper might not be reliable enough to be put to use by law enforcement or other authorities anytime soon, but it does provides crucial gifts to future academics to work on. One of the main issues faced by researchers while training machine learning models is that they can't ever find abundant quality databases to train them on. But, this paper provides the researchers with the gift of two different databases to train the algorithms on to carry out similar tasks, each having a whopping 2,000 images.

The algorithm described in the system can successfully identify faces in disguises like sunglasses, bandanas, scarves, masks, but it does face issues when it comes to identifying faces disguised in rigid Guy Fawkes masks, which are often donned by members of hacking collective Anonymous. The researchers are currently working on ways they can even ID people disguised in these rigid marks. What's impressive to note is that, the experimental algorithms are already successful in identifying people with 99 per cent accuracy on the basis of how they walk.

The researchers have big things planned for the future. They want their algorithm to work in real time without a Wi-Fi connection. This way, people using the tech would be able to use a camera to ID people wearing disguises without having to manually feed data into a remote served over a Wi-Fi. The team also plans on training the system on a larger dataset of people to sharpen its accuracy.

While on the face of it, the idea might seem brand new, but it actually isn't. A paper published in 2014 also discussed an automated algorithm that could recognise faces even when they were obscured. The new paper makes uses of a different technique and a larger data set. Prior to that, a 2008 paper analysed the effect disguises have on facial recognition algorithms.

Interestingly, in order to restore the balance, there's a plenty of research also being done on how to successfully evade these facial recognition algorithms. For instance, last year, a team at Carnegie Mellon University developed glasses that could fool a facial recognition algorithm.

Tata Elxsi Invites Tech Startups For Its Newly Launch Incubation Program

Tata Elxsi, a leading global design and technology services company, is inviting startups having deep technology based product, platform, or solutions for Indian consumers in India to their Incub@TE program. The Tata Elxsi incub@TE program has been set up to foster entrepreneurial activities in the technology space. This enables the startups associated with the program to get better visibility and an association that brings trust in their target market.

The program will be for 9 months, providing a platform for startups and prospective entrepreneurs to convert their ideas into commercially viable businesses.

For these startups, incub@TE has tied up with 1Crowd to support funding needs and Nasscom 10K as ecosystem partner in order to further help these startups, foster entrepreneurial activities in the deep technology space. Tata Elxsi’s incub@TE will be able to take the startups’ ideas to the market by providing ready-to-use infrastructure, key technical consulting through in-house experts, mentorship, access to service providers, plug-n-play office and connections with people and organizations that matter.

Speaking on the development, Rajesh Kumar, Vice President - Strategic Initiatives, Tata Elxsi said, “Deep technology is a very futuristic concept, but is slowly taking lead in India, and globally. We, at Tata Elxsi, want to enhance this growth, and see many potential startups in the country that can help contribute to this progress. We look forward to onboarding some impactful startups working on deep technologies, and provide them with the right infrastructure and guidance, in order to help bring this development to the forefront.”

Relevant startups working on deep technology will have to present their ideas to the incub@TE team at Tata Elxsi, with information such as market validation, founders and key team member’s details, business offering along with target market, positioning, and current status.

Tata Elxsi, a part of the $ 100 billion Tata group is amongst the world’s leading providers of design and technology services for product engineering and solutions across industries including Broadcast, Communications, and Automotive. It provides technology consulting, new product design, development, and testing services. Tata Elxsi also provides solutions and services for emerging technologies such as IoT (Internet of Things), Big Data Analytics, Cloud, Mobility, Virtual Reality and Artificial Intelligence. This is backed by over 25 years of engineering experience, a global delivery presence and offshore development centres in India.