AI Pe Charcha is a series of panel discussions initiated by MeitY (Ministry of Electronics and Information Technology). It’s a monthly series that has been conceptualized following the success of “Raise 2020 Global Summit,” which was inaugurated by the honorable Prime Minister on October 5, 2020. Most recently, the talk in AI Pe Charcha was about AI for Data Driven Governance.
AI Pe Charcha involves discussions by several global and domestic leaders from Government and Industry, sharing their views and experiences on artificial intelligence-related case studies, best AI practices around the globe, and challenges faced in the field of artificial intelligence. As a part of the series, MeitY organized AI Pe Charcha on the theme of “AI for Data Driven Governance”.
The session was held on October 28, 2021, where panel members discussed leveraging AI for the public sector, postal services, future smart cities, defense, and security. It includes the importance of data-driven AI services that are accepted globally. The panel had industry experts like Abhishek Singh (CEO of MyGov and NeGd), Dr. Martin Klein (Global GM of SAP), and Rahul Lodhe (Senior Director of SAP Artificial Intelligence).
The core topics discussed were, why the Indian Government needs data-driven governance, including intelligent technologies like AI? How do the US, Australia, and European governments use data-driven AI Governance to improve citizen services? How does the SAP framework help the government in the strategic use of Data?
The session had presentations on ‘AI for the public sector creating a Digitally Intelligent Data’ presented by Dr. Martin Klein and ‘Possibilities with AI for Data Driven governance’ presented by Rahul Lodhe.
With AI Pe Charcha on AI for Data Driven Governance, the Indian Government paves the way to understand and witness the power of emerging technologies to the common people.
Meta AI formerly known as Facebook, collaborated with researchers at Carnegie Mellon to develop a kind of synthetic skin named ReSkin that can be used in robotic hands. Researchers mention that it is pretty challenging to develop touch-sensitive technologies for artificial intelligence and robotics.
Meta has created a unique robotic skin that has the capability of assisting artificial intelligence-powered robots in retrieving various information like object weight, texture, temperature, and state.
It is a cost-effective and replaceable technology that uses machine learning algorithms, soft robotics, and magnetic sensing technology to generate critical object information. According to Meta, the technology can be effortlessly integrated with existing artificial intelligence-powered robots.
Postdoctoral researcher at Meta, Tess Hellebrekers, said, “I want this sensing skin to be so robust and simple that anyone could set it up and start gathering touch data within a day.”
The newly developed artificial skin will play a vital role in advanced artificial intelligence-based robotics research and projects as it solves one of the most critical challenges of touch sensing and collecting rich contact data.
Meta mentioned in a blog that ReSkin could be used to perform various touch-based tasks, including object classification, proprioception, and robotic grasping. Mark Zuckerberg said, “We designed a high-res touch sensor and worked with Carnegie Mellon to create a thin robot skin. This brings us one step closer to realistic virtual objects and physical interactions in the metaverse.”
Meta and Carnegie Mellon researchers tested ReSkin on a robot to touch, lift, and grab soft fruits like grapes, strawberries, and blueberries. The test showed excellent results and displayed one-of-a-kind capabilities of the newly developed technology.
Apart from ReSkin, Meta also unveiled its new humanoid finger robot named DIGIT that is capable of understanding human-level touch in the physical world using tactile sensing.
DeepMind, an artificial intelligence subsidiary of Alphabet Inc., is widely known for its contribution to deep reinforcement learning, specifically expertise in complicated games and predicting protein structures. Now, the organization aims to take the next step in robotics research by acquiring MuJoCo.
A recent update on the DeepMinds website describes acquiring the rigid-body physics simulator MuJoCo and has made it freely available for the research community. MuJoCo (Multi-Joint Dynamics with Contact) is a physics engine that aims to facilitate research and development in robotics, biomechanics, graphics and animation, and other areas where quick and accurate simulation is required. As robotics research is freely available, it will have a progressive impact on the work of scientists who are struggling with the costs. At a glance, MuJoCo becomes a vital factor for DeepMinds future. It not only serves as a science or artificial intelligence lab but also serves a business unit for one of the largest tech companies in the world.
Training and testing robots in the practical world is an expensive and slow process. Simulation platforms are a big boon for research in robotics, it allows researchers to train multiple AI agents in parallel at a faster speed than in real life. Today, with advancements in simulated environments, most robotics research teams carry out the bulk of training in their AI models. A model, once trained, enters a testing phase to fine-tune real physical robots further.
In the past few years, several simulation environments for reinforcement learning and robotics have been launched. However, MuJoCo stands out from many physics simulators like Isaac Gym, PyBullet, or Roboschool, because it has fine-grained detail that simulates contact surfaces. It accurately models the physics laws shown in the emergence of physical phenomena such as Newton’s Cradle.
MuJoCo also has an admirable feature that supports simulation of the musculoskeletal model of humans and animals, which is principally crucial in bipedal and quadruped robots. Often, the differences in a model result in performance degradation of AI models. The increased accuracy of the physics environment will ensure minor deviation when comparing results of simulated environments to the real world, also called the ‘sim2real’ gap. A smaller sim2real gap denotes lesser changes when implemented in the physical world.
Before DeepMind open-source MuJuCo, many researchers had to bear the license cost and rely on the PyBullet platform. A recent paper by researchers in Cardiff University states that “the annual cost of MuJoCo institutional license is $3000,” which is an unreasonable deal for many small research teams when a project has a more extended time period. OpenAI released Roboschool in 2017, which was an alternative to MuJoCo (for Gym), its toolkit to train deep reinforcement learning models for robotics and various applications.
DeepMinds also refers to a recent article in PNAS (proceedings of the national academy of the sciences) emphasizing the application of simulation in robotics. The authors have recommended better support for the development of an open-source simulation platform, mentioning “a robust and feature-rich set of four or five simulation tools in the open-source domain is critical to have state-of-the-art solutions in robotics.” To align with the proposed goal, DeepMind has committed to develop and maintain MuJoCo as a free, open-source, community-driven platform. It is also important to remember that MuJoCo is designed to leverage power only from traditional processors (CPUs), not GPUs with many computation cores.
In contrast to the interest, Researchers at the University of Toronto, Nvidia, and other organizations highlighted the limits of MuJoCo simulation platforms that work on CPUs. For instance, OpenAI developed a robotic hand called Dactyl, which is trained on a computational cluster of around 30,000 CPU cores. These costs remain a challenge with CPU-based platforms.
Richard Sutton is one of the pioneers of reinforcement learning and describes it as “the first computational theory of intelligence.” According to DeepMind’s scientists, if one has a complex environment and a sound reinforcement learning algorithm, one can develop AI agents that will acquire the traits of general intelligence. DeepMind envisions developing artificial general intelligence (AGI) that is flexible to understand innate problem-solving capabilities found in humans and animals.
While scientists debate for the ideal paths to AGI, DeepMind has demonstrated its views timely. MuJoCo acquisition can provide DeepMind a powerful tool to test this hypothesis and gradually gain better results. By availing it to small research teams, DeepMind can also help garner talent it would hire in the future.
Intel has launched the Intel Unnati program to set up 100 Unnati data-centric labs in emerging technologies across all engineering institutions in India. This initiative will assist in equipping engineering students across the country with relevant industry skills.
The Unnati program will establish 100 data-centric laboratories across the nation to improve research infrastructure and technology excellence among students.
The idea from the chipmaker brings academy and industry together for contributing to the industry-needed skills. Focused on recent technologies, the Unnati program allows learning new things like emerging technologies with their use cases and a hands-on approach to techniques via Unnati labs.
Intel will serve its part as the technology and knowledge partner, while the engineering institutions should bear the expenses of the lab. Institutions will be provided with the choice of lab variants like training the faculty and providing maintenance support according to their budgets.
As a part of the program, subscribed institutions will be provided with hardware and software recommendations, course content, and course completion certificates for students. At present, the Unnati data-centric labs are available for artificial intelligence (AI), AI Internet of Things (AIIoT), FPGA solutions, and other recent technologies that include smart mobility and intelligent security. The Intel Unnati program had a trial run in 15 colleges that are in the process of establishing the data-centric labs on their campuses. Three of these labs are currently in use.
“Intel Unnati data-centric labs will enable educational institutions to bridge the widening technology skill gap in the country, build industry-ready emerging technology competencies and provide strategic impetus to India’s digital economy transformation,” said Nivruti Rai, country head of Intel India and vice president of Intel Foundry Services.
To enable a more significant focus on innovation and research, Intel has committed to expanding digital promptness to reach 3 crore people in 30,000 plus institutions in more than 30 countries by 2030.
Tesla, an American electric vehicle and a clean energy-based company, has released a new whitepaper regarding a new standard for the Dojo supercomputing platform. This standard specifies arithmetic formats and methods for arithmetic (floating point) operation in computer programming environments during deep learning neural network training.
Tesla handles an insane amount of video data from its fleet of over 1 million vehicles to train its neural nets. Over the last two years, Tesla has been teasing the development of a new in-house supercomputer called ‘Dojo’ that optimizes neural net video training.
Automakers found themselves unsatisfied with current hardware options to train their computer vision neural nets and believed they could do better internally. CEO Elon Musk has marked the paper as “more important than it may seem.” The dojo was recently unveiled by Tesla at Tesla’s AI Day in August, and it could potentially become the most powerful supercomputer in the world.
The whitepaper on Dojo technology primarily aims to define a standard that provides a method for computing floating-point numbers. This method yields similar results whether the processing is done in hardware, software, or a combination of both.
Developers initially got motivated from the original IEEE 754 standard (1985), that specified formats and methods for floating-point arithmetic in computer systems. Recently, Google Brain, an artificial intelligence research group at Google, developed the Brain Floating Point (BFloat16) format in their TPU (tensor processing unit) architecture for their machine learning training systems. The BFloat16 format is consumed by some major processors like Intel, ARM, AMD, TensorFlow, or Nvidia. It still differs from IEEE Float16 format in the number of bits allocated for mantissa and exponent bits as shown below:
Comparing IEEE and Brain 16 bit Floating point formats. Image Source: Tesla Dojo Whitepaper
Tesla extended precision support by introducing Configurable Float8 (CFloat8), an eight-bit floating-point format. This format reduces memory storage and bandwidth in storing weights, activations, and gradient values essential for training and increasing larger networks. While IEEE Float16 and Bfloat16 format have a fixed number of bits allocated to mantissa and exponent bits, eight bits can only accommodate a small number of mantissa and exponent bits. CFloat8 requires some configurability to ensure high accuracy and convergence of the training models.
Even last year, Musk teased that Tesla’s Dojo would have a capacity of over an exaflop, which means quintillion (1018 ) floating-point operations per second. Although the automaker already has a Dojo chip and tile, it is still building its full rack to create the supercomputer.
The biopharma and healthcare industry across the globe are facing immense challenges for the wrath of the COVID-19 pandemic. With a decrease in consumer affordability and significant price pressure of companies, it is becoming quite a challenge for pharma and healthcare companies to meet their sales target and increase revenue generation.
According to a report, over 30% of health patients in the United States are not consuming prescribed doses of medicine due to the rising cost. Accenture has come up with a strategic plan that mentions the need to improve the research and development capabilities of organizations to tackle the global issue.
The research mentions, “The productivity problem hinges on the rising ratio of R&D spend per each new treatment approved, which has increased at 5% per NME and 7% per approval annually over the last ten years.”
Accenture claims that an increase in investment for research and development can drastically cut down costs of bringing discoveries from billions to millions. Deployment of better analytics and cutting-edge technologies in research and development can significantly boost innovation and productivity.
Accenture suggests pharmaceutical companies put more emphasis on three different prospects, namely New Science portfolio, digital and data-led research, and faster, smarter development, in order to bring a change in the situation.
The use of artificial intelligence solutions can help reduce the research cost by targeting validation, lead optimization, and identification. It would also play a vital role in cutting down costs due to failures due to a higher Probability of technical and regulatory success (PTRS).
Companies can save over $1.2 billion for every successful medicine and also generate additional revenue of up to $450 million after the adoption of Accenture’s business strategy. Modernizing the R&D process of pharmaceutical and healthcare companies can help them get out of this profitability crunch caused by the COVID-19 pandemic.
Technology giant Amazon partners with the University of Carolina, Los Angeles (UCLA), to launch a new research center for dealing with the rising number of social issues caused due to the amplification of artificial intelligence technologies across the globe.
The newly established center named ‘The Science Hub for Humanity and Artificial Intelligence’ is Amazon’s first-ever collaboration with a public university to offer academic support in areas of mutual interest in the field of artificial intelligence and how it can be deployed to benefit humanity.
Amazon has planned to provide funding of $1 million in the first year of this partnership to support research projects and doctoral fellowships on the University’s campus. Chancellor of UCLA, Gene Block, said, “The Science Hub for Humanity and Artificial Intelligence will advance AI-related discoveries and deepen our understanding of a discipline that is revolutionizing the way we use and understand modern technology.”
He further added that they are thrilled to partner with Amazon in carving the future of artificial intelligence and its applications globally. The science hub aims to understand the broadly faced challenges in the artificial intelligence industry and develop solutions to tackle them that would be beneficial for the entire society.
Vice President of Applied Science at Amazon Web Services AI, Stefano Soatto, said, “The hub is designed to foster the educational mission of the university, so it can best educate the diverse talent needed to sustain the AI revolution in the years to come, in a way that benefits all sectors of society.”
He also mentioned that the science hub would create new opportunities for university researchers by showing them various challenges and issues that Amazon faces while carrying out its operations. The science hub funding will also provide financial assistance to students of the second, third, and fourth-year by offering annual fellowship for living and tuition fees.
KLA Corporation joins hands with the Indian Institute of Technology (IIT) Madras to launch its new artificial intelligence-powered advanced computing lab at the IIT Madras Research Park. KLA will be opening two new facilities in Chennai, Tamil Nadu, as a step towards the firm’s goal of investing in research and development and creating new talents in the country.
According to officials, the facilities will focus on research in the field of artificial intelligence. The company develops artificial intelligent process control and process enabling solutions for the semiconductor manufacturing industry at a global scale. The new research centers will help the company further grow its operations and hire new employees from different regions of the country.
President of semiconductor process control at KLA, Ahmed Khan, said, “Our researchers and engineers at AI-ACL join the AI experts at our AI Modeling and Center of Excellence in Michigan to form a global team committed to advancing the boundaries of AI, software, image processing, and physics modeling.”
He further added that KLA is the industry-leading in using artificial intelligence solutions to identify and isolate challenges in chip manufacturing, and the new AI ACL research center would help them expand their reach of artificial intelligence in product development.
The company currently hosts over 1200 employees, and with the new facilities, new hirings can be expected. Earlier this year, KLA Corporation had raised a pandemic relief fund of over $550,000 to support the country’s COVID-19 crisis by providing financial aid to various healthcare facilities.
Director of IIT Madras, Prof Bhaskar Ramamurthy, said, “The IIT Madras Research Park ecosystem is a perfect enabler for such an industry with academic collaboration that is bringing together our resident experts, top student researchers, and industry’s best minds. I also congratulate KLA on the grand opening of its new office in RMZ Millenia-II today.”
He also mentioned that KLA and IIT Madras have been collaborating for the past 15 years, and this new association will help them accelerate research in parallel computing and quantum computing to develop new artificial intelligence-enabled applications for the semiconductor and metrology industries.
Intel AI Global Impact Festival has bestowed Mehreen Mushtaq Shamim, a teacher from Delhi Public School, with an award in artificial intelligence. She got selected as one of the four winners in the ‘AI impact shapers: Teachers with innovative AI teaching learning practices’ program.
Teachers from over 20 countries participated in this global competition. Teachers from South Korea, Singapore, and Poland won the other three awards. Since the inclusion of artificial intelligence in the CBSE curriculum, Mehreen has been teaching the subject in DPS to students of grades 9th to 12th. The subject is now available from grade 6th in CBSE schools.
Mehreen said, “I introduced students to Bootcamps that give them advanced lessons in AI. I could mentor students and inspire them into creating projects, three of which are now pending patents. Our students are interested in technology and can think out of the box very well. If guided in the right direction, they can do wonders.”
To date, Mehreen has impacted many lives, including empowering over 1200 students and training 250 teachers in artificial intelligence. Mehreen has completed her Masters in Computer Applications and nurtured more interest in artificial intelligence as she kept on learning more about the subject to teach her students.
“Knowing the power of AI, I now want to encourage children of other streams — not just science — to take up the subject and upskill themselves,” she said. The Intel event began on 15th October 2021 and hosted over 200 artificial intelligence-powered social impact innovations and recognized them at a global level.
Samsung-owned healthcare company Neurologica receives Food and Drug Administration (FDA) clearance for its new artificial intelligence technology named Auto Lung Nodule Detection (ALND) that detects lung nodules in chest X-rays.
The new technology provides on-device computer-assisted detection of pulmonary nodules from 10mm to 30mm in size. The solution uses artificial intelligence algorithms to generate reports.
This technology will drastically help doctors and healthcare professionals to diagnose and identify issues and take quick measures while treating patients. The solution has already been tested in several academic centers, and it shows a sensitivity of over 80% for detecting pulmonary lung nodules.
Vice President of digital radiography and ultrasound at Neurologica, David Legg, said, “This FDA clearance is a huge milestone for Samsung and is the result of our tireless work to design diagnostic solutions that empower providers to deliver patients the absolute best care possible.”
He further added that the technology’s accuracy and reliability would allow doctors to present it in front of patients with utmost confidence. The technology also comes with an autorun option that automatically starts processing images after X-ray and offers a PACS transmission option to suit hospital environments. However, the tool cannot be used on patients with lung lesions other than abnormal nodules.
United States-based medical imaging equipment manufacturing firm Neurologica was founded by Ibrahim Bechwati in 2004. The company had raised over $17 million in two funding rounds before being acquired by Samsung in 2013. Neurologica specializes in developing healthcare equipment that helps doctors and hospitals to provide better patient experience, patient care, patient satisfaction and also optimize hospital workflow.
