Third IBM IEEE CAS/EDS AI Compute Symposium (AICS’2020)
The 3rd IBM IEEE CAS/EDS AI Compute Symposium, known as (AICS’20), was held over two days on Oct 21-22, 2020. The symposium was also an initiative supported by IBM Academy of Technology (https://www.ibm.com/blogs/academy-of-technology/). Dr. Joshi has been the main interface for CAS and EDS for organizing this successful event. The symposium was very well attended and received positive responses from the audience. It was organized as a virtual symposium first time due to the pandemic situation across the globe. More than 650 folks from academia, industry, research organizations including 380 IEEE members registered from over 45 countries. The salient features of the symposium included high quality keynotes and invited talks, over 40 student posters, best paper poster awards, and excellent panel discussions. The theme of the symposium was “From Atoms to Applications”.
Keynote talks were delivered by Dr. Mukesh V. Khare (VP IBM Research), Prof. Rob Knight (Director and Professor at UC San Diego) and Pamela Norton (CEO of Borsetta).
Dr. Mukesh V. Khare presented an exciting vision of “What's Next in AI: Our Vision for the Future of AI Hardware.” The talk highlighted the theme that we have reached a turning point in computation. As the global pandemic and warming climate present challenges, the next generation of computers will define and respond to these and future crises. The IBM Research AI Hardware Center is focusing on the development of a new, scalable computing platform that uses the power of AI and the flexibility of the hybrid cloud to create a virtually limitless pool of computing power and capabilities that can be applied to the challenges we face.
Prof. Rob Knight followed up with a keynote “Using AI to Understand the Human Microbiome and its Role in COVID-19.” The human microbiome harbors an immensely complicated collection of genes and cells that outnumber “our” own and are redefining the concept of what it means to be human. Grappling with data of this complexity requires artificial intelligence techniques. Highlighting the collaboration between IBM and UCSD in the Artificial Intelligence for Healthy Living, he showed applications of cutting-edge techniques to understand the microbiome, discovering how it changes with age and disease to recommend modifications to diet, lifestyle and medications that can increase health throughout the lifespan. Of particular interest during today’s unprecedented COVID-19 pandemic are studies regarding how the microbiome affects those who are the most impacted by SARS-CoV-2, including the elderly and those with specific microbiome-lined co-morbidites. These techniques will be generally applicable to other illnesses and will help us bridge the previously disparate fields of infectious and chronic disease.
Following Prof. Knight’s keynote, Prof. Deji Akinwande (Endowed Professor Electrical and Computer Engineering University of Texas, Austin) described how atomic level research can expand the applications of AI. His talk entitled “Atomic Memory: From Single Defects to Analog Switches and Computing,” focused on the discovery of the memory effect in 2D atomically-thin nanomaterials, providing a greater scientific understanding and advancements for engineering applications. Non-volatile memory devices based on 2D materials are an application of defects and is a rapidly advancing field with rich physics that can be attributed to vacancies combined with metal diffusion. In particular, the talk highlighted his group’s pioneering work on monolayer memory (atomristors) that could enable various applications including zero-power devices, non-volatile RF switches, and memristors for neuromorphic computing.
Next, Prof. James J. DiCarlo, MD, PhD (Professor, Co-director at Quest for Intelligence, Department head, investigator of Institute for Brain Research, M.I.T) enlightened the audience with “Reverse Engineering Visual Intelligence.”
The brain and cognitive sciences are hard at work on a great scientific quest — to reverse engineer the human mind and its intelligent behavior. Yet these fields are still in their infancy. Not surprisingly, forward engineering approaches that aim to emulate human intelligence (HI) in artificial systems (AI) are also still in their infancy. Yet the intelligence and cognitive flexibility apparent in human behavior is an existence proof that machines can be constructed to emulate and work alongside the human mind.
The talk predicted that these challenges of reverse engineering human intelligence would be solved by tightly combining the efforts of brain and cognitive scientists (hypothesis generation and data acquisition), and forward engineering aiming to emulate intelligent behavior (hypothesis instantiation and data prediction). As this approach discovers the correct neural network models, those models will not only encapsulate our understanding of complex brain systems, they will be the basis of next-generation computing and novel brain interfaces for therapeutic and augmentation goals (e.g., brain disorders).
Further, the talk focused on one aspect of human intelligence — visual object categorization and detection — and how work in brain science, cognitive science and computer science converged to create deep neural networks that could support such tasks. These networks not only reach human performance for many images, but their internal workings are modeled after— and largely explain and predict — the internal workings of the primate visual system. Yet, the primate visual system (HI) still outperforms current generation artificial deep neural networks (AI), suggesting that the brain and cognitive sciences can offer new clues.
These recent successes and related work suggest that the brain and cognitive sciences community is poised to embrace a powerful new research paradigm. More broadly, our species is at the beginning of its most important science quest — the quest to understand human intelligence. In short, his talk motivated the audience to engage that frontier alongside us.
The following speaker, Dr. Laurens van der Maaten (Research Director at Facebook AI Research in New York) talked about “Exploring the Limits of Weakly Supervised Pretraining.” State-of-the-art visual perception models for a wide range of tasks rely on supervised pretraining. ImageNet classification is the de facto pretraining task for these models. Yet, ImageNet is now nearly ten years old and is by modern standards “small”. Even so, relatively little is known about the behavior of pretraining with datasets that are multiple orders of magnitude larger. The reasons are obvious: such datasets are difficult to collect and annotate. In this talk covered a unique study of transfer learning with large convolutional networks trained to predict hashtags on billions of social media images. The experimental data demonstrated that training for large-scale hashtag prediction led to excellent results. Improvements on several image classification and object detection tasks were shown and the highest ImageNet-1k single-crop, top-1 accuracy to date: 85.4% (97.6% top-5) reported. Extensive experimentation provided novel empirical data on the relationship between large-scale pretraining and transfer learning performance.
On the second day of the symposium Pamela Norton (CEO, Borsetta) gave a keynote on importance of security, entitled “Securing the Future of AI on the Edge with Intelligent Trusted Chips.” To usher in this new edge economy, we need to ensure that the convergence of AI, new compute processing and intelligent chips migrating to the edge will protect our privacy and security. This keynote explored a framework to protect our privacy, security, and create new opportunities for all in this emerging new economy.
The following speaker, Prof. Dheerisha Kudithipudi (Director of the MATRIX AI Consortium and the Robert F McDermott Endowed Chair in Engineering and Professor in ECE & CS University of Texas, San Antonio) gave a talk entitled “Neuro-Inspired AI: Compact and resilient models for the Edge.” Neural plasticity offers insights into designing artificial intelligent systems. As the convergence of these two fields is in a nascent phase, there is a critical knowledge gap in designing neuromorphic systems that can support on-device learning with heterogeneous plasticity. She presented a recent in-silico learning system from the Nu.AI lab based on a
CMOS/Memristor architecture. It emulated a biomimetic sequence memory algorithm inspired by the neocortex with structural and intrinsic plasticity mechanisms. The salient feature included a synthetic synapse representation that supports dynamic synaptic pathways for compact memory. The structural plasticity in the synaptic pathways was emulated in the memristor device’s physical behavior and the synaptic modulation is achieved through a custom training scheme.
Jay Gambetta (IBM Fellow and VP of Quantum Computing, IBM) next presented a quantum roadmap in a talk entitled “Quantum circuits, and the Future of Quantum Technology in the Cloud.” In the past few years, quantum computing has moved beyond the laboratory setting and has been accelerated through cloud access. It is a new kind of computing that uses the same physical rules that atoms follow to manipulate information. At this very fundamental level, quantum computers execute quantum circuits much like a computer executes logic circuits, but by using the physical phenomena of superposition, entanglement, and interference, to implement mathematical calculations that are out of the reach of even the most advanced supercomputers. He gave an overview of the IBM quantum effort to increase the device performance of superconducting qubit systems to produce quantum circuits with higher fidelity and how IBM is linking the computational difficulty of these circuits towards quantum applications. This is an exciting era where cutting edge research, system and software development are pushing the frontier forward, bolstered by an engaged and growing quantum computing community.
In the final talk of day two Prof. Laurent Daudet (CTO and Co-Founder, LightOn; Professor at Universite de Paris) presented “Unlocking Transformative AI with Photonic Computing.” Recent large-scale AI models, such as OpenAI’s GPT-3 for NLP, open a wide range of economic opportunities, with possibly even more impact than deep learning had in the last decade. However, training these models requires massive amounts of computing power. This talk covered LightOn’s view on how future AI hardware should be designed for addressing some of the hardest computing challenges, such as in natural language processing (NLP), recommender systems, and big science. In particular, the talk highlighted how LightOn’s Optical Processing Units
(OPUs) can be seamlessly integrated into a variety of hybrid photonics / silicon pipelines implementing state-of-the-art Machine Learning algorithms.
The symposium also featured a poster session. Out of 44 poster papers, 4 best poster papers were awarded. The list of winners was as follows:
- Tianyu Jia, Yuhao Ju, Russ Joseph and Jie Gu,” NCPU: A Binary Neural Network that Emulate RISC-V CPU at the Conjunction of Neuromorphic and Von-Neumann Architectures”, Northwestern University, Evanston, IL
- Seah Kim and Hasan Genc, “Gemmini: Enabling Systematic Deep-Learning Architecture Evaluation via Full-Stack Integration”, University of California, Berkeley, CA
- Sanjeev T. Chandrasekaran, Sumukh Bhanushali, Imon Banerjee and Arindam Sanyal, “Towards intelligent wearable health monitors using reservoir computing CMOS IC”, State University of New York at Buffalo, NY
- Minh Truong, ”Data-Oriented Processing-Using-Memory with Emerging Memory Technology”, Carnegie Mellon University, Pittsburgh, PA
The symposium closed with a panel discussion entitled “What will be the currency of future AI computation?” The panel, moderated by Dr. Arvind Kumar, consisted of Pamela Norton, Prof. Dheerisha Kudithipudi, Dr. Sidney Tsai, Prof. Laurent Daudet, Prof. James J. DiCarlo, and Dr. Kristan Temme, representing expertise in security, analog AI, optical computing, brain science, and quantum computing. Each panelist gave an initial point of view on the panel topic, followed by questions from the audience. The initial points of view suggested benefits for each approach, but there was a general consensus that no one approach would dominate, and heterogenous computing with secure translation between ‘currencies’ would be needed. Technical questions were posed on a range of topics, such as the combination of approaches, emulation of Hebbian learning, plasticity and its importance, the longevity of current DNNs, and the role of Quantum Computing in AI. The panel ended with some questions from the audience seeking advice from the panelists for early researchers in the field.
Special thanks to Cindy Goldberg (IBM) and Brittian Parkinson (CAS) for all the publicity help.
Please visit the symposium webpage for more information about this year’s symposium and future editions. https://www.zurich.ibm.com/thinklab/AIcomputesymposium.html
Dr. Rajiv Joshi (General Chair)
Dr. Arvind Kumar (Program co-Chair), Dr. Matt Ziegler (Program co-Chair)
Dr. Xin Zhang (Poster co-Chair), Dr. Krishnan Kailas (Poster co-Chair)
________________
Women in Circuits and Systems (WiCAS) and Young Professionals (YP) at ICECS 2020
A “Women in Circuits and Systems” (WiCAS) event and a “Young Professionals” (YP) event took place during the 27th IEEE International Conference on Electronics Circuits and Systems (ICECS), the flagship conference for the Region 8 of the IEEE Circuits and Systems Society (CASS).
The WiCAS events traditionally aim to inspire and motivate both students and young professionals in the domain of circuits and systems to have efficient roles in their professions, by meeting successful female engineers and professors, through interesting technical and professional talks in fields of interest of CASS. The YP events usually include start-ups presentation, poster and demo sessions, aiming to provide a thrilling environment for early career researchers to present their work. Joining this event, young professionals have the opportunity to learn about state of the art and most advanced activities in the area of circuits and systems, meet and interact with their peers, receive feedback from internationally well-known experts in the CASS domain from both academia and industry.
The ICECS 2020 WiCAS event was held on November 24th, 2020, organized by Dr. Erika Covi (NaMLab, Germany), Dr. Giulia Di Capua (University of Salerno, Italy), and Dr. Melika Payvand (ETHZ, Switzerland).
Figure 1 (right): Some attendees of the WiCAS session
The WiCAS event included a panel discussion, driven by the theme “Inspiring stories of women's success in STEM”. The event involved the participation of four recognized speakers, both from academia and industry: Prof. Teresa Serrano-Gotarredona (National Microelectronics Institute, Seville, Spain), Dr. Angeliki Pantazi (Research Manager at IBM, Zurich, Switzerland), Dr. Chiara Bartolozzi (Italian Institute of Technology, Genoa, Italy), and Dr. Elham Ebrahimi (Comms Technical Lead at Spire, Edinburgh, United Kingdom). As women in leadership positions with proven excellent record of accomplishment in fields of interest of the IEEE CASS, the WiCAS panelists shared precious experiences and advices with the conference attendees. The speakers not only shared their personal journey and the path to their success, but also their research interest. Their research and experience spanned from analog/mixed-signal design and neuromorphic engineering, all the way to emerging technologies and even space applications. The panel discussion was a lively session on the impact of gender gap and cognitive biases in women’s career.
The ICECS 2020 YP event was held on November 25th, 2020, organized by Dr. Nazila Fough (Robert Gordon University, United Kingdom) and Dr. Giulia Di Capua (University of Salerno, Italy).
Figure 2 (left): Some attendees of the YP session.
The YP event included welcoming talks about IEEE, IEEE YP and IEEE CAS, generously provided by Dr. Mona Ghassemian (IEEE UK & Ireland section chair), Dr. Vinko Lešić (IEEE YP Region 8 chair), Prof. Izzet Kale (IEEE UK & Ireland vice-chair and CAS Chapter-chair) and Prof. Tony Davies (IEEE R8 History Activities coordinator). A discussion session on how to get research works successfully published was also organized, thanks to the amazing contributions of Dr. Hakim Meskine (Advance Electronics Materials Editor in Chief, Wiley) and Dr. Charles Glaser (Springer Nature Editorial Director). After these talks, a YP demo session and a YP poster session took place. The YP track was open to students, new lecturers and early career researchers, who had the opportunity to present their research work, stay technically up-to-date and learn about future challenges in CASS domain.
During the conference, awards were assigned for the best contributions presented in the WiCAS and YP competitions. All WiCAS and YP winners were announced during the ICECS 2020 awarding ceremony.
Twenty-seven papers were accepted for the WiCAS competition, and three awards were assigned. A “1st Runner-Up Paper Award” was assigned to Dr. Saman Fatima, a Doctoral Researcher at the University of Paris Saclay (France), for the paper titled “A CMOS Readout Circuit for a Low Shunt Resistance IR photo-Detector”. A “2nd Runner-Up Paper Award” was assigned to Ms. Eve McGlynn, a PhD Student at the University of Glasgow (United Kingdom), for the paper titled “Encapsulated Magnetoelectric Composites for Wirelessly Powered Brain Implantable Devices”. A “Best Paper Award” was eventually assigned to Dr. Fanny Spagnolo, a Postdoctoral Researcher at the University of Calabria (Italy), for the paper titled “A High-Performance and Power-Efficient SIMD Convolution Engine for FPGAs”.
Twelve papers and four demo presentations were accepted for the YP competition, and three YP awards were eventually assigned. A “Best Demo Award” was assigned to Dr. Jason Eshraghian, a Post-Doctoral Fellow at the University of Michigan (U.S.A.), for the live demonstration titled “Prosthesis Control Using a Real-Time Retina Cell Network Simulator”. A “Best Presentation Award” was assigned to Ms. Elena-Diana Șandru, a PhD student at the University Polytechnic of Bucharest (Romania), for the poster titled “Machine Learning-Based Local Sensitivity Analysis of Integrated Circuits to Process Variations”. Finally, a “Best Paper Award” was assigned to Mr. Kosuke Uchiyama, a PhD student at the Nagoya University (Japan), for the poster titled “Design of Fully-Integrated Self-Powered FM Transmitter Using On-Chip Photodiodes in 65-nm CMOS”.
The WiCAS and YP winners received a one-year IEEE plus IEEE CASS membership, thanks to the generous support provided by the IEEE CASS. Moreover, as a sign of appreciation to all 2020 WiCAS and YP participants, the WiCAS co-chairs and the YP co-Chairs were pleased to grant a 25 USD voucher for CAS-related technical books to all speakers who gave live presentations during the conference days.
The WiCAS and YP co-chairs express their gratitude to the ICECS 2020 general co-chairs, Prof. Hadi Heidari and Prof. Elena Blockina, and to the conference organizing committee, for the technical support received in the definition of the WiCAS and YP sessions. The WiCAS and YP co-chairs would also like to thank the IEEE Circuits and Systems Society for the financial support that helped making these events a great moment of professional and personal enrichment.
Erika Covi, Giulia Di Capua, Nazila Fough and Melika Payvand