|
Sunday
|
Monday
|
Tuesday
|
Wednesday
|
Thursday
|
Friday
|
Saturday
|
|---|---|---|---|---|---|---|
|
0 events,
|
0 events,
|
2 events,
-
Team theory is a mathematical formalism for decentralized stochastic control problems in which a “team,” consisting of a number of members, cooperates to achieve a common objective. It was developed to provide a rigorous mathematical framework of cooperating members in which all members have the same objective yet different information. In static team problems, the […]
-
Large language models (LLMs) have advanced the frontiers of AI reasoning: they can synthesize information from multiple sources, derive new conclusions, and explain those conclusions to their users. However, LLMs do not do this reliably. They hallucinate facts, convincingly state incorrect deductions, and exhibit logical fallacies like confirmation bias. In this talk, I will describe […] |
0 events,
|
2 events,
-
The remarkable recent success of Machine Learning (ML) is driven by our ability to develop and deploy interactive models that can solve complicated tasks by understanding and adapting to the ever-changing state of the world. However, the development of such models demands significant data and computing resources. Moreover, as these models increasingly interact with humans, […]
-
Promises are cheap. Software vendors routinely describe their offerings as “secure”, but few are based on designs that can guarantee even the most basic security properties. To address this problem, services like Cloudflare, Android, and Firefox are increasingly relying on languages like Rust and WebAssembly to provide safety by design. But these promises too can […] |
2 events,
-
The rapid adoption of generative AI has created a cycle where personal information cascades perpetually: from people to models to applications and online platforms, then back through scrapers into the system. Simple blanket rules such as "don't train on this data" or "don't share sensitive information" are inadequate, as we face training data scarcity while […]
-
Special location for this talk: 105 Amy Gutmann Hall Recent advances in Artificial Intelligence are powering revolutionary interactive tools that will transform the very nature of the scientific enterprise. We describe several large-scale projects at the Allen Institute for AI aimed at developing open models, agentic platforms, and novel interaction paradigms in order to amplify […] |
0 events,
|
|
0 events,
|
0 events,
|
0 events,
|
1 event,
-
Abstract: Extracting insights from imaging data used to be straightforward: every component of imaging systems was engineered by humans, the analysis and interpretation of the collected data was driven by human understanding and experience, and only humans were responsible for the impact of the decisions stemming from such insights. Today, however, machine learning permeates every […] |
2 events,
-
Zoom link https://upenn.zoom.us/j/98220304722 Abstract We present computational tools for analyzing and designing first-order methods in parametric convex optimization. These methods are popular for their low per-iteration cost and warm-starting capabilities. However, precisely quantifying the number of iterations required to compute high-quality solutions remains a key challenge, especially in real-time applications. First, we introduce a […]
-
We have made exciting progress in AI by massive models on massive amounts of data center compute. However, the demands for AI are rapidly expanding. I identify how to maximize performance under any compute constraint, expanding the Pareto frontier of AI capabilities. This talk builds up to an efficient language model architecture that expands […] |
1 event,
-
Modern datacenters must handle an ever-growing array of real-time and data-intensive workloads, such as interactive web services and AI models, that demand both low latency and high throughput. However, traditional operating systems introduce significant I/O overhead, degrading performance and reducing efficiency. A common solution is to let applications directly communicate with hardware, bypassing the operating […] |
0 events,
|
|
0 events,
|
0 events,
|
0 events,
|
0 events,
|
0 events,
|
0 events,
|
0 events,
|
|
0 events,
|
0 events,
|
1 event,
-
Generative modeling offers a powerful paradigm for designing novel functional DNA, RNA and protein sequences. In this talk, I introduce probabilistic experimental design methods to efficiently manufacture samples from generative models of biomolecules in the real world. These algorithms merge computational techniques for approximate sampling with physical randomness. I also develop tools to rigorously evaluate […] |
1 event,
-
Abstract: Data-driven systems hold immense potential to positively impact society, but their reliability remains a challenge. Their outputs are often too brittle to changes in their training data, leaving them vulnerable to data poisoning attacks, prone to leaking sensitive information, or susceptible to overfitting. Establishing fundamental principles for designing algorithms that are both stable—to mitigate these […] |
2 events,
-
Zoom link: https://upenn.zoom.us/j/98220304722 Abstract: Modern ML models produce valuable predictions across various applications, influencing people’s lives, opportunities, and scientific advancements. However, these systems can fail in unexpected ways, generating unreliable inferences and perpetuating biases present in the data. These issues are particularly troubling in high-stakes applications, where models are trained on increasingly diverse, incomplete, and […]
-
Artificial intelligence (AI) is transforming scientific discovery, particularly in materials science, by accelerating the prediction and design of materials with desired properties. Traditional physics-based modeling of atomic systems is computationally prohibitive for large-scale problems, and AI addresses this challenge by learning the underlying physics from data, thereby accelerating discoveries. In this talk I will present […] |
0 events,
|
0 events,
|
|
0 events,
|
0 events,
|
1 event,
-
Cyber-physical systems (CPS), powered by emerging artificial intelligence (AI) technologies, have become integral to various critical domains such as the Internet of Things (IoTs), medical devices, and autonomous vehicles. A unique aspect of these systems lies in their interactions with the physical world, by perceiving environments through heterogeneous modalities (perception), processing digital data with intelligence […] |
1 event,
-
Abstract: Controlling language models is key to unlocking their full potential and making them useful for downstream tasks. Successfully deploying these models often requires both task-specific customization and rigorous auditing of their behavior. In this talk, I will begin by introducing a customization method called Prefix-Tuning, which adapts language models by updating only 0.1% of […] |
2 events,
-
Zoom link: https://upenn.zoom.us/j/98220304722 Abstract: We study the generalization properties of neural networks through the lens of data complexity. Recent work by Buzaglo et al. (2024) shows that random (nearly) interpolating networks generalize, provided there is a small ``teacher'' network that achieves small excess risk. We give a short single-sample PAC-Bayes proof of this result and […]
-
Many domains of machine learning, from language modeling to computer vision, have recently undergone a shift towards generalist models, whose broad generalization abilities are fueled by large and diverse real-world training datasets and high-capacity model architectures. In robotics, however, it has been challenging to apply the same recipe: after all, we cannot easily scrape millions […] |
3 events,
-
Perceiving the 4D world (i.e., 3D space over time) from visual input is essential for human interaction with the physical environment. While computer vision has made remarkable progress in 3D scene understanding, much of it remains piecemeal—for example, focusing solely on static scenes or specific categories of dynamic objects. How can we model diverse dynamic […]
-
We have made exciting progress in AI by massive models on massive amounts of data center compute. However, the demands for AI are rapidly expanding. I identify how to maximize performance under any compute constraint, expanding the Pareto frontier of AI capabilities. This talk builds up to an efficient language model architecture that expands […]
-
The Cora Ingrum Center for Community and Outreach is planning its annual Celebration of Community gala to showcase Penn Engineering students, staff, and faculty in their multi-talented richness. The event will consist of guest speakers, performances, presentations from student groups, and a variety of cuisines. Do not hesitate to contact Dr. Ocek Eke (ocek@seas.upenn.edu) and […] |
0 events,
|
|
0 events,
|
0 events,
|
1 event,
-
Generative models are revolutionizing our world, with the ability to generate photorealistic visual content that are indistinguishable from reality. Despite their overwhelming presence in the cyber world, they haven’t been very useful in the physical world that we live in. In this talk, I will present how the rich priors learned by large-scale generative models—ranging […] |
1 event,
-
Abstract: Machine learning applications are increasingly reliant on black-box pretrained models. To ensure safe use of these models, techniques such as unlearning, guardrails, and watermarking have been proposed to curb model behavior and audit usage. Unfortunately, while these post-hoc approaches give positive safety ‘vibes’ when evaluated in isolation, our work shows that existing techniques are quite brittle when deployed […] |
2 events,
-
Zoom link: https://upenn.zoom.us/j/98220304722 Abstract I will present recent work by my research group on the design and analysis of stochastic-gradient-based algorithms for solving nonconvex constrained optimization problems, which may arise, for example, in informed supervised learning. I will focus in particular on algorithmic strategies that have consistently been shown to exhibit the best practical […]
-
I aim to build complete intelligent agents that can continually learn, reason, and plan: answer queries, infer human intentions, and make long-horizon plans spanning hours to days. In this talk, I will describe a general learning and reasoning framework based on neuro-symbolic concepts. Drawing inspiration from theories and studies in cognitive science, neuro-symbolic concepts serve […] |
1 event,
-
We have made exciting progress in AI by massive models on massive amounts of data center compute. However, the demands for AI are rapidly expanding. I identify how to maximize performance under any compute constraint, expanding the Pareto frontier of AI capabilities. This talk builds up to an efficient language model architecture that expands […] |
0 events,
|
