Abstract:
Coarse grained (CG) models alleviate the drawbacks of all-atom simulations. The latter still pose challenges because they are computationally expensive and give access to limited spatiotemporal scales, despite the use of modern high-performance computing clusters. CG models ignore some of the atomistic degrees of freedom, leading to fewer interatomic interactions, hence less computing time. Introducing such models emphasizes the need to properly manage these multiple scales, by carefully deriving potentials and reconstructing conformations from their CG representations, usually with the help of Machine Learning. Following a bottom-up and force matching approach, we train a Physics-Informed Neural Network to extract the CG force field parameters from all-atom simulation data. We verify our approach by applying it to fibrin monomers to study multiple-fibrin polymerization in solution at the microsecond scale, after modifying the force field to incorporate further non-bonded interactions, not present in the training data. Access to these scales will allow us to study the effects of some of the molecules' components. Furthermore, we modify recent solutions in data-driven protein backmapping. Taking advantage of the developments in graph neural networks and variational inference, we introduce an intermediate step in the all-atom reconstruction of a molecule given its CG configuration, in an attempt to more accurately de-coarsen structures whose atom-to-CG-beads ratio is very high. The combined effect of our new forward and inverse coarse graining methodology will enable the in silico study of many phenomena that are highly dynamic and intrinsically multiscale.

Bio:
Georgios Kementzidis is a third year PhD student in the Department of Applied Mathematics and Statistics at Stony Brook University. His advisor is Dr. Yuefan Deng. His research interests lie at the intersection of Computational Science, molecular dynamics (MD) simulations, and Machine Learning (ML) applications to Computational Biophysics. He is particularly interested in coarse-graining and multi-scale simulations.

*Note: this seminar will be held in-person (food provided on a first-come, first serve basis) and online*

Join Zoom Meeting https://stonybrook.zoom.us/j/99510099036?pwd=EyowuLBGvUVLZDBlG6F6chkMICFOZ7.1
Meeting ID: 995 1009 9036
Passcode: 132419
Abstract: Large Language Model (LLM) agents have demonstrated remarkable generalization capabilities across multi-domain tasks. Existing agent tuning approaches typically employ supervised finetuning on entire expert trajectories. However, behavior-cloning of full trajectories can introduce expert bias and weaken generalization to states not covered by the expert data. Additionally, critical steps--such as planning, complex reasoning for intermediate subtasks, and strategic decision-making--are essential to success in agent tasks, so learning these steps is the key to improving LLM agents. For more effective and efficient agent tuning, we propose ATLAS that identifies the critical steps in expert trajectories and finetunes LLMs solely on these steps with reduced costs. By steering the training's focus to a few critical steps, our method mitigates the risk of overfitting entire trajectories and promotes generalization across different environments and tasks. In extensive experiments, an LLM finetuned on only 30% critical steps selected by ATLAS outperforms the LLM finetuned on all steps and recent open-source LLM agents. ATLAS maintains and improves base LLM skills as generalist agents interacting with diverse environments.

Speaker: Bijoy

Location: Old Computer Science Building - CS2311
Join CELT on Tuesday, March 31 for a focused, one-hour overview on how to redesign and future-proof assessments in the age of AI! This session will cover three key areas: leveraging AI as a co-pilot for developing effective exam questions, designing authentic assessments, and exploring how AI can strategically support active learning structures like Team-Based Learning (TBL), Project-Based Learning (PBL), and Scenario-Based Learning (SBL).

Register here.
Abstract: Gaussian Probability Path-based Generative Models (GPPGMs) generate data by reversing a stochastic process that progressively corrupts samples with Gaussian noise. Despite state-of-the-art results in 3D molecular generation, their deployment is hindered by the high cost of long generative trajectories, often requiring hundreds to thousands of steps during training and sampling. In this work, we propose a principled method, named GAGA, to improve generation efficiency without sacrificing training granularity or inference fidelity of GPPGMs. Our key insight is that different data modalities obtain sufficient Gaussianity at markedly different steps during the forward process. Based on this observation, we analytically identify a characteristic step at which molecular data attains sufficient Gaussianity, after which the trajectory can be replaced by a closed-form Gaussian approximation. Unlike existing accelerators that coarsen or reformulate trajectories, our approach preserves full-resolution learning dynamics while avoiding redundant transport through truncated distributional states. Experiments on 3D molecular generation benchmarks demonstrate that our GAGA achieves substantial improvement on both generation quality and computational efficiency.

Speaker: Jingxiang Qu

Location: New Computer Science 220
Educational objectives:

1. Explain how AI represents a Cognitive Revolution in academic medicine, redefining thefundamental limits of human cognition and knowledge work.
2. Differentiate between superficial AI adoption (innovation theatre) and truetransformationthrough AI-native institutional design.
3. Analyze how AI fundamentally reshapes clinical, research, and educational work-from dataentry to verification, recall to recognition, and hypothesis generation to evaluation-andidentify implications for redesigning academic health systems.

Speaker: Jiajie Zhang, Ph.D.,Dean, Professor, and Glassell Family FoundationDistinguished Chair in Informatics Excellence,D. Bradley McWilliams School of BiomedicalInformatics, UTHealth Houston

Location: MART Building, Room: 7M-0602 (7th Floor)
Abstract: The rapid growth of observational data presents unprecedented opportunities to enhance both the predictability and mechanistic understanding of Earth systems. However, fully harnessing big Earth data needs computational frameworks that bridge the gap between physics-based models and machine learning. In this talk, I will first demonstrate how AI methods can significantly improve the prediction of environmental systems. Despite their predictive accuracy, machine learning models often lack physical interpretability, limiting their ability for scientific inquiry. To address this, I will introduce the developed hybrid, differentiable modeling framework that unifies physical models with machine learning in an end-to-end trainable system. This framework autonomously learns from large observations while maintaining physical clarity. The machine learning components can be seamlessly embedded into physical backbones to assimilate multi-source data, support automatic parameterization, and represent uncertain processes. I will showcase applications of this framework in simulating and understanding the terrestrial water cycle and its interactions with ecosystems at continental and global scales. This talk will highlight how differentiable modeling not only improves the modeling ability in both data-rich and data-scarce scenarios, but also provides a systematic pathway to enhancing model structures, deciphering uncertain physical relations, and facilitating knowledge discovery in Earth system sciences.


IACS Seminar Speaker: Dapeng Feng, Stanford Univeristy

Location: IACS Seminar Room
Abstract: Visual generation is a fundamental problem in computer vision and graphics, with applications ranging from 3D capture to content creation and image/video synthesis. Despite rapid progress in neural rendering and generative models, efficiency remains a key obstacle in practice: high-quality 3D reconstruction often depends on dense multi-view supervision; scalable 3D synthesis faces heavy optimization, training, and rendering costs; and modern image/video generators incur substantial computation as token grids grow with spatial resolution and temporal length.
This thesis targets efficient visual world modeling by improving sample efficiency in 3D reconstruction, representation efficiency in 3D generation, and computational efficiency in image/video synthesis. First, we improve sample efficiency for neural implicit surface reconstruction under sparse views by integrating multi-view stereo probability volumes as a geometric regularizer, enabling high-quality reconstruction from as few as three input images. Next, we introduce an explicit 3D representation for 3D generation, built from multi-view depth and RGB predictions with 3D Gaussian features, which enables the use of 2D generative priors while enforcing multi-view consistency via epipolar attention. We then address the computational bottleneck of image and video synthesis with importance-based token merging, using importance signals available during generation to preserve critical information while merging redundant tokens. Finally, we propose efficient mixed-resolution diffusion transformers via cross-resolution phase-aligned attention, aiming to improve attention stability under mixed token grids and support high-fidelity mixed-resolution generation.

Speaker: Haoyu Wu

Location: NCS120

Fantastic Futures is an international conference series organized by AI4LAM as one of the most important global events at the intersection of AI and cultural heritage, shaping how memory institutions adapt to rapidly evolving technologies while maintaining public trust.

The conference brings together professionals, researchers, technologists, and cultural‑heritage institutions to explore how AI can support organizational workflows, users services and data preservation, access, discovery, as well as innovation in the cultural‑heritage sector:

  • Librarians, archivists, museum professionals and all other GLAM enthusiasts
  • Technologists working on cultural‑heritage applications, AI researchers and developers
  • Digital humanists, researchers of cultural heritage
  • University professors, students, and librarians
  • Policy and ethics experts, governmental employees
  • Strategic leaders and innovation managers

To register, visit AI4LAM's official website.