Abstract: For decades, we have built software by writing code, but in recent years machine learning has emerged as a new approach to create software with features that would be impossible to code by hand. However, the use of learning to build software risks ignoring some of the lessons we have learned for how to […]
Abstract: Neuroimaging has significantly expanded our understanding of brain changes in neuropsychiatric disorders as well as in aging and neurodegenerative diseases. However, it wasn’t until the advent of machine learning tools that imaging signatures that can be detected in individuals, rather than groups, were constructed. More importantly, imaging signatures derived via machine learning models have […]
Abstract: The explosive growth of machine learning and data-driven methodologies have revolutionized numerous fields. Yet, the translation of these successes to the domain of dynamical physical systems remains a significant challenge. Closing the loop from data to actions in these systems faces many difficulties, stemming from the need for sample efficiency and computational feasibility, along […]
The ASSET and Warren Center will be hosting a research mixer to welcome new PhD students and expose new and current students to the wide range of research in Penn Engineering in AI/ML. The program will consist of (short) faculty talks, poster presentations by students/postdocs, and a reception to end the night!
Abstract: Despite efforts to align large language models (LLMs) with human intentions, popular LLMs such as chatGPT, Llama, Claude, and Gemini are susceptible to jailbreaking attacks, wherein an adversary fools a targeted LLM into generating objectionable content. For this reason, interest has grown in improving the robustness of LLMs against such attacks. In this talk, we review the current state of […]
The solutions to many of society’s most pressing problems rely on the discovery of materials with unprecedented physical and chemical properties that are tailored to an application of interest. Typically, it is not a matter of incremental improvements over existing technologies; rather, there is often an urgent need to identify new kinds of materials altogether. […]
Abstract: Large pre-trained models trained on internet-scale data are often not ready for safe deployment out-of-the-box. They are heavily fine-tuned and aligned using large quantities of human preference data, usually elicited using pairwise comparisons. While aligning an AI/ML model to human preferences or values, it is important to ask whose preference and values we are […]
