Each year, the Intel® Rising Star Faculty Award (RSA) program selects early-career academic researchers who are leading the advancements in technology research that demonstrate the potential to disrupt the industry. The 15 award winners chosen for 2023 are being recognized for their novel works in computer science, electrical engineering, computer engineering, and chemical engineering.
The program proudly recognizes community members who are doing exceptional work in the field and hopes to facilitate long-term collaborative relationships with senior technical leaders at Intel. Award recipients are also chosen for their innovative teaching methods and for their efforts to improve the fields of computer science and engineering.
The research conducted by the selected faculty provides novel solutions to challenges spanning various topics, including computer architecture, software systems, quantum computing, AI for autonomous systems, graph neural networks, 3D fabrication, deep neural networks, and robotics.
This year’s winners consist of faculty members from the following institutions:
- Carnegie Mellon University
- Cornell University
- Georgia Institute of Technology
- Indian Institute of Science – Bengaluru, India
- Stanford University
- Tel Aviv University
- The Ohio State University
- Trinity College Dublin
- University of California, Los Angeles
- University of California, San Diego
- University of Illinois Urbana-Champaign
- University of Pennsylvania
- University of Southern California
- University of Texas at Austin
- University of Washington
The 2023 RSA Winners Are:
Intel’s Rising Star Faculty Award Winners for 2023 (from top row, left): Anthony Sigillito, Bolei Zhou, Akshitha Sriraman, Dan Congreve, Mahmood Sharif, Callie Hao, Owolabi Legunsen, Saugata Ghose, Larisa Florea, Shamsul Arafin, Simon Du, Swabha Swayamdipta, Utsav Banerjee, Xiaolong Wang, and Yaoyao Jia.
Anthony Sigillito
Assistant Professor, Electrical and Systems Engineering
University of Pennsylvania
Professor Sigillito’s group at Penn develops Si/SiGe-based quantum devices. In particular, his group is focused on developing novel quantum control approaches and device architectures with the goal of improving device scalability and connectivity. These include manipulating spins through excited states and developing a new and creative shuttling architecture based on resistive top-gated interconnects. Additionally, the Sigillito lab has developed classical control hardware tailored to qubit devices that operate at both cryogenic and room temperatures. These include voltage sources with better-than-commercial noise performance, ultra-low-noise cryogenic HEMT transimpedance amplifiers, and all-superconducting microwave phase and amplitude modulators to potentially enable loss-less cryogenic frequency conversion.
Bolei Zhou
Assistant Professor, Computer Science
University of California, Los Angeles
Zhou’s research is at the intersection of computer vision and machine autonomy, focusing on developing interpretable and generalizable embodied AI for autonomous systems. The rapid advancement of autonomous systems driven by AI has transformed various industries, from autonomous vehicles reshaping transportation to warehouse robots automating packaging processes. However, the black-box nature of deep neural networks in these systems raises significant AI safety and accountability concerns, as they can pose significant risks to the surroundings if not operated properly. To address this, Zhou aims to harness machine learning and simulation to enhance autonomous system safety before real-world deployment. Zhou's research goal is to create a data-driven approach for learning to simulate diverse realistic scenarios and behaviors to assess AI safety, benefiting various domains such as autonomous driving, home robots, and warehouse automation.
Akshitha Sriraman
Assistant Professor, Electrical and Computer Engineering
Carnegie Mellon University
Akshitha Sriraman’s research bridges computer architecture and software systems, with a focus on making hyperscale data center systems more efficient, sustainable, and equitable via solutions that span the systems stack. Her work has developed the software and hardware foundations of hyperscale data center systems that support modern web services, such as web search, video streaming, and online healthcare. Her past work on systematically designing data center hardware for the hyperscale era by efficiently using existing hardware and carefully designing new hardware, i.e., accelerators, has improved the efficiency of real-world data centers that serve billions of users, saving millions of dollars, and meaningfully reducing the global carbon footprint. Using her prior research as a supporting pillar, Sriraman is defining the next generation of research that tackles the daunting problem of building efficient web systems in a way that is also socially responsible through bold, holistic system design approaches. She is introducing cross-stack solutions to improve the efficiency of web systems in a socially responsible manner by improving (1) sustainability, to reduce web systems’ effects on anthropogenic climate change, and (2) equity, to identify and minimize inequities (e.g., demographic biases) in existing web systems and to design new web systems that elevate the lives of historically underserved rural communities. In addition to her research, Sriraman is dedicated to increasing the participation of underserved students in science and engineering and has developed strategies to increase the ratios of students from underserved communities and ensure their retention.
Assistant Professor, Electrical Engineering
Stanford University
Professor Congreve engineers nanoscale systems to solve challenging problems, with a particular focus on materials and devices to control light and energy. Using these nanomaterials, Congreve is developing next generation volumetric 3D printers that could be used for advanced manufacturing of batteries, photonic systems, 3D lithography, and other high-throughput, high-resolution applications, building on his research recently published in Nature. He further aims to develop next generation light emitters utilizing perovskite nanomaterials, both in LED and lasing modalities, which could open new designs in sensing and communications, and he is working to further understand how these material systems could be used in devices including photovoltaics, transistors, and memory. Finally, he is investigating how photon energy conversion could drive innovations by increasing the spectral window for sensing, biological, cryptographic, and photochemical applications.
Assistant Professor, Computer Science
Tel Aviv University
Machine-learning (ML) algorithms enable transformative technologies (e.g., face recognition, cancer diagnosis, and autonomous driving), and are becoming increasingly ubiquitous in computer systems. Still, introducing ML algorithms into systems may potentially cause harm. For example, ML may increase the attack surface of computer systems. Indeed, adversarial examples – slightly but strategically perturbed variants of benign inputs to mislead ML models – pose a profound challenge to ML integrity. The deployment of ML also raises critical concerns related to data privacy, the extent to which we can explain systems’ decisions, and potential discriminatory treatment of individuals. Sharif’s work studies the trustworthiness of ML systems from multiple angles, with the aim of enabling trustworthy deployment of ML, such that society would be able to reap the benefits of ML while minimizing harm.
Assistant Professor, Electrical and Computer Engineering
Georgia Institute of Technology
Professor Hao's groundbreaking work on the very first Deep Neural Network (DNN) and Accelerator co-design methodology (DAC’19) has had a significant impact on the research community. This research has been followed by several subsequent efforts, establishing the concept of true simultaneous co-design by merging algorithm and accelerator design spaces. Furthermore, Dr. Hao's research encompasses a comprehensive and practical study of general Graph Neural Network (GNN) and dynamic GNN architecture design and automated compilation. Hao’s studies mark pioneering efforts in the field, with the added distinction of providing publicly available source code that has been validated on actual FPGAs. Her proposed GNN architecture has garnered significant interest from scientists in High Energy Physics (HEP), who require a nano-speed GNN accelerator on FPGA for their particle detection experiments.
Assistant Professor, Computer Science
Cornell University
Owolabi Legunsen works on runtime verification, software testing, and the unification of these techniques, towards transforming how software quality assurance is done today. He pioneered the research on integrating runtime verification with regression testing, creating exciting connections between testing and formal verification. Legunsen also invented two types of tests to fill decades-old gaps in software testing practice: (1) cTests validate cloud-system configuration changes against code to prevent costly production-time failures caused by misconfigurations, and (2) inline tests check for bugs in individual program statements. Legunsen's work on configuration testing is building an intellectual bridge between software engineering and systems research. His research has helped discover hundreds of confirmed bugs in scores of open-source projects and his inline testing tool is now part of Python's testing framework. Over the next few years, Legunsen plans to develop new algorithms and tools to specialize runtime verification for software testing, so that more developers can check program executions against formal specifications. A main challenge that is being addressed is how to democratize runtime verification among software engineers without requiring developers to be formal methods experts.
Assistant Professor, Computer Science
University of Illinois Urbana-Champaign
Professor Ghose’s work is making significant progress towards bringing widespread processing-in-memory (PIM) closer to reality. PIM addresses the high energy and latency costs of moving data between the CPU and memory, by enabling computational capabilities near or inside memory cells. While research on PIM has been going on for decades, and while initial PIM prototypes are now commercially available, there remains a large gap between the narrow solutions that these prototypes provide and the potential promise of PIM across a wide range of application domains. Professor Ghose’s work aims to bridge this gap, by focusing on PIM programmability and completing the first full stack for general-purpose PIM systems. To date, his work has addressed a broad set of challenges, from automating the identification of PIM program kernels, to co-designing architectures with emerging memory devices to deliver a practical, scalable approach to performing digital logic directly inside memory cells. He leads the ARCANA Research Group, where his students are working on multiple directions to improve the intelligence of memory and storage systems, and to develop the next generation of data-centric computer systems.
Associate Professor, Chemistry
Trinity College Dublin
Larisa Florea is an Associate Professor in Chemistry and Materials Science in School of Chemistry & AMBER, the SFI Research Centre for Advanced Materials and BioEngineering Research, Trinity College Dublin. Since establishing her group in 2018, Professor Florea has become a recognized leader in the field of soft robotics. Her unique approach is enabled by her background in organic chemistry and polymer science, and her expertise in 3D fabrication and micro-fabrication technologies. Florea’s research tackles some of the grand challenges in materials research, by showing that intelligent chemistry, smart design, and precise engineering can give enhanced capabilities to soft materials. Her research combines fundamental material synthesis that provides access to ‘adaptive’ materials that switch between different characteristics in response to various external stimuli. Stimulation of these materials using light, temperature, electrochemical potential, or changes in the local chemical environment can result in highly precise 4D control, from the nano to the macro scale. The incorporation of responsive units at the molecular level, combined with precise assembly at the nano-/micro- level via 3D micro-fabrication technologies enables controllable functions such as movement, transport, and sensing to be given to inanimate materials, enabling the creation of autonomous soft robots. The goal of Professor Florea’s research is to augment the capabilities of traditional robotic tools, through the use of responsive bio-inspired polymeric constructs, to revolutionize their operation at the micro-scale.
Assistant Professor, Electrical and Computer Engineering
The Ohio State University
Professor Arafin has made significant contributions in a number of areas. His research and educational focus primarily lie in both electrical engineering and material sciences. In particular, his areas of focus include III-V compound semiconductor and 2D-materials, optoelectronic devices, semiconductor lasers and photonic integrated circuits which have potential to significantly advance semiconductor technologies and the future of computing. Since he formed the Optics and Photonics Research Lab (OPREL) 4 years ago, his research group has performed cutting-edge research on a wide range of areas including quantum dot based narrow-linewidth widely tunable lasers, mid-wave infrared photonic integrated circuits, as well as quantum materials and the associated single photon sources. He has also extended his experimental research in the area of topological photonics given its potential and promise to offer many attractive properties such as, scatter-free edge-state transport, immune to perturbations and disorder to the associated active devices.
Assistant Professor, Computer Science and Engineering
University of Washington
For several years now, deep learning has achieved unprecedented success in a number of domains, including understanding scenes and natural language. However, we have little understanding of the underlying principles of why deep learning works. In a series of works, Professor Du resolved two of the most important open questions in explaining the success of deep learning: optimization and generalization. One of the biggest mysteries in deep learning is how stochastic gradient descent (SGD) is able to find the global minima of deep neural network training. Although the objective function is highly non-convex and non-smooth, Du gave the first proof of global convergence of SGD in deep neural network training. The second mystery is how deep learning can still generalize well despite the fact that the model size is much larger than the data size. In a seminal paper, Professor Du established a foundational and innovative equivalence between deep learning and another mature machine learning method, kernel learning; this connection explains, among other things, why deep learning generalizes well.
Assistant Professor, Computer Science
University of Southern California
Professor Swayamdipta’s research involves designing algorithms to select, investigate and synthesize data for training language models aimed at enhancing the capabilities of these models. While there has been an immense emphasis on scaling up the training resources for language models, there has been little attention given to actual knowledge content in these resources. Swayamdipta’s research is aimed at disrupting these current practices through the creation of measures of usable information in the data, so training language models can both be efficient and result in higher capabilities. She also designs algorithms for better generative capabilities of the language models, which enable knowledge retrieval from smaller-scale models, presenting a formidable alternative to large-scale language models which are not accessible to a wider audience due to resource constraints. Finally, her research also discovers societal biases in the existing training data resources which pose the risk of being enhanced and propagated in the models that consume them. She has built algorithms that are aimed at mitigating such risks through data-centric approaches involving selection and incorporation of contextual knowledge that are specifically aimed at enhanced social cognition in language models.
Assistant Professor, Department of Electronic Systems Engineering
Indian Institute of Science – Bengaluru, India
Professor Banerjee leads the Secure Intelligent and Efficient Systems (SINESys) Lab. His research interests include cryptography, hardware security, digital circuits, embedded systems and chip design. Security is a major concern in the context of Internet of Things (IoT) and edge computing. Most IoT devices are resource-constrained embedded systems, which makes it difficult to implement computationally expensive cryptographic algorithms such as post-quantum cryptography, fully homomorphic encryption, functional encryption, zero-knowledge proofs and secure machine learning. These implementations must also be secured against side-channel attacks to prevent the leakage of sensitive information. Professor Banerjee's research addresses these challenges through novel hardware-based security solutions – design of efficient and configurable cryptographic accelerators, low-overhead side-channel countermeasures and co-design of custom hardware with standard micro-processors such as RISC-V.
Assistant Professor, Electrical and Computer Engineering
University of California, San Diego
Professor Wang’s group at UC San Diego has conducted research mainly in the directions of Computer Vision, Machine Learning and Robotics. His research covers a wide range of work from learning foundation models with large-scale data to on-device efficient and robust deployment of these models on real robots (e.g., legged robot navigation, dexterous manipulation). On the computer vision side, his work focuses on learning 3D and dynamics representations through videos and physical robotic interaction data. His group explores various means of training signals using self-supervision, language, and common-sense knowledge. On the robotics side, his group leverages these comprehensive representations to facilitate the learning of robot skills, with the goal of generalizing the robot to interact effectively and efficiently with a wide range of objects and environments in the real physical world. The intersection of software development on large-scale learning and the co-development of hardware and embedding systems for real-robot deployment symbols the unique strength of Professor Wang’s group.
Assistant Professor, Electrical Engineering
University of Texas at Austin
Professor Jia's research endeavors have resulted in many low-power, high-performance analog/mixed signal integrated circuit designs, particularly sensing frontends, data converters, and power converters. These innovative designs not only exhibit exceptional power efficiency, operating at levels in the nanowatt and microwatt ranges, but also push the boundaries of tradeoffs among power, noise, resolution, and other crucial factors. With their innovative features, these circuits hold great potential to find widespread applications in smart homes, autonomous vehicles, healthcare wearables, and other IoT devices. Beyond integrated circuit design, Professor Jia's research extends to the system level, leading to the development of powerful research devices that can overcome the limitations of current tools, paving the way for discoveries in neuroscience. These devices carry immense transformational potential for clinical applications spanning from closed-loop brain-machine interfaces to the treatment of neurological diseases and conditions such as epilepsy, Alzheimer’s disease, and Parkinson’s disease.