Amazon ML Fellows for 2022-2023
USC and Amazon have created a joint research center focused on development of new approaches to machine learning (ML) privacy, security, and trustworthiness. The Center for Secure and Trusted Machine Learning (in short, Trusted AI), which will be housed at the USC Viterbi School of Engineering, will support USC and Amazon researchers in the development of novel approaches to privacy-preserving ML solutions.
SINA SHAHAM
SINA SHAHAM
Biography
Sina Shaham is currently a third-year Ph.D. candidate within the computer science department at USC, advised by Prof. Cyrus Shahabi. His research is on the interaction of Privacy and Fairness in Machine Learning and its application for Geospatial data. Previously, he was an intern at Meta working on fairness in social media and has held positions as a Researcher and Data Scientist in the industry. He has over a dozen publications on enhancing Privacy and Fairness in decision-making.
Research Interests
My research aims to study the relationship between the two common pillars of human ethics, i.e., Privacy and Fairness in Machine Learning. Coming from an Information Theory background, I tend first to identify the impact of one on another, quantify the interaction between two objective functions and ultimately propose mitigation techniques that consider both objectives. The primary application I am focused on is Geospatial datasets due to their significance in elections and fund allocation.
What does being an Amazon ML Fellow mean to you?
Amazon has a long standing reputation for technological innovation and support for creativity. Being an Amazon fellow provides a unique opportunity for me to explore new ideas and contribute to the journey of Responsible AI in shaping the future.
YUNHAO GE
YUNHAO GE
Biography
Yunhao Ge is currently a third-year Ph.D. student in the Computer Science Department of University of Southern California. He is advised by Prof. Laurent Itti. His research interests are in Machine Learning, Computer vision, and their applications towards more Human-centric, Humanoid and Trustworthy AI.
Research Interests
My research interests are in Machine Learning, Computer vision, and their applications towards more Human-centric and Humanoid AI. My current research focuses include:
(1) Human-centric properties of AI models (Causal Explainable AI, Human-to-AI / AI-to-AI Knowledge Exchange, Domain Adaptation, Out-of-distribution Detection (OOD))
(2) Simulate human cognitive learning ability (Continual Learning, Imagination, Reasoning, Visual Recognition)
(3) How generative models (NeRF, GAN, VAE) and multi-modal models (DALL-E, CLIP) help downstream discriminative models (detection, segmentation)
What does being an Amazon ML Fellow mean to you?
Being an Amazon fellow means not only an honor but motivation for me. It encourages me to keep exploring the challenging task of understanding the reasoning logic of AI models, and further simulating the human's cognitive learning ability toward more Human-centric, Humanoid, and Trustworthy AI.
JIAO SUN
JIAO SUN
Biography
Jiao Sun is a Computer Science Ph.D. candidate at the University of Southern California. She is advised by Professor Xuezhe (Max) Ma, and collaborates closely with Professor Nanyun (Violet) Peng and Swabha Swayamdipta. Her works on trustworthy Natural Language Generation have appeared in top NLP and HCI conferences, including a best paper recommendation at ACL 2021 and a best paper honorable mention at CHI 2022. She did research internships at Amazon Alexa AI, Google Research and IBM Research in the past. Before coming to USC, she finished her master's at Tsinghua University. Jiao focuses on trustworthy Natural Language Generation (NLG) during her Ph.D. More specifically, she builds controlled text generation models and enhances the robustness of NLG systems. Along this line, she also works on proposing better text generation evaluation metrics and investigating fairness problems. When there is a need to understand human perception, Jiao never hesitates to bring humans in the loop to better understand the utility of NLG systems.
What does being an Amazon ML Fellow mean to you?
Words cannot express how excited I am to be an Amazon ML fellow! It is a great recognition and encouragement for me to continue working on trustworthy text generation. We have always been surprised about fluent and human-like text that models could generate, but we should enhance the control to produce more reliable and fair content. With the fellowship, I hope we can push one step further towards improving and applying state-of-the-art text generation models in fruitful use cases that Amazon could provide!