The Association for Computational Linguistics is the international scientific and professional society for people working on problems involving natural language and computation. Membership includes the ACL quarterly journals, Computational Linguistics and Transactions of the ACL, reduced registration at most ACL-sponsored conferences, discounts on ACL-sponsored publications, and participation in ACL Special Interest Groups.

An annual meeting is held each summer in locations where significant computational linguistics research is carried out.

For more information and registration, visit the official website.

Talk by Zhenhua Liu to be followed by AI Institute updates


Abstract: Decision making with uncertainty has been studied in multiple communities extensively. Recently, online optimization has gained popularity partially because of its promising performance guarantees by incorporating predictions. In this talk, I will provide an overview of our work on algorithm designs for online optimization and its applications. Then, I will talk about our recent work in ACM Sigmetrics 2019 on choosing predictions and control algorithms simultaneously and dynamically. Finally, I will discuss some ongoing efforts and collaboration opportunities.

Bio: Zhenhua Liu is currently an assistant professor in the Department of Applied Mathematics and Statistics at Stony Brook University. He is also affiliated with the Department of Computer Science, the AI Institute and the Smart Energy Technology Cluster. He received his PhD degree in Computer Science from California Institute of Technology. His current research interests include cloud computing, online optimization and learning, smart grid, market design and distributed control. His research combines rigorous analysis and system design, and goes from theory, to prototype, and eventually to industry to make real impacts.

You are cordially invited to attend the biweekly Brookhaven AI Mixer (BAM). BAM includes three short talks on AI research happening at BNL, followed by an open mixer over coffee and snacks for everyone to network and discuss all things AI. The first half hour will consist of presentations that will be available via ZOOM, and the second half hour will be for in person only networking.

Join us every other Tuesday at noon in CDSD's Training Room (building 725, 2nd floor) to learn about interesting AI methods and applications, engage with potential collaborators, prepare for pending FASST funding calls, and build a community of AI for Science at BNL.

Tuesday, November 26, 2024, 12:00 pm -- CDS, Bldg. 725, Training Room

Speakers

Hanfei Yan, NSLS-II

David Park, CDS, AI Dept

Xihaier Luo, CDS, AI Dept

Join Zoom Meeting

https://bnl.zoomgov.com/j/1601052863?pwd=eIX9qZKPGNtQ11uwbK8JP5hIdIxA3V.1

Meeting ID: 160 105 2863

Passcode: 442980


The talk will be exclusively on zoom https://stonybrook.zoom.us/j/7851507944 Speaker: Sooyeon Lee, Rochester Institute of Technology Title: Design and Evaluation of Accessible AI Technologies for Users with Disabilities Abstract: Over one billion people in the world live with some type of disability. Many of them experience barriers in accessing information or using technologies, which can limit social interactions in both physical and digital spaces. In my research, I focus on investigating and designing nonvisual interaction for the community of blind users and non-audio and non-speech interaction for the community of deaf and hard of hearing users. In this talk, I will first present my research investigating nonvisual interaction prototypes for supporting shopping activities for blind users, with an exploration of one-way instructional and two-way conversational interactions and with a variety of form factors and communication modalities through the use of human-computer interaction research methodologies. I will also discuss incorporation of AI technology and its impact on the nonvisual guidance experiences, and further meanings of independence and new ways for designing independence for people with visual impairments. This collaborative work included AI researchers, the community of the blind, and an industry research partner. Additionally, I will discuss my findings and further exciting research opportunities. Secondly, I will overview research projects investigating AI-based applications and tools that support deaf and hard of hearing people's equitable information access and societal participation. This work addresses engagement in online social media spaces, workplace communication, participation in gig work, and interaction with mainstream technology through American Sign Language (ASL) interaction. I will focus on a recent project on users' experiences with AI deep-fake face-transformation technologies to support anonymous participation of deaf and hard of hearing signers in online social media. Lastly, I will discuss my future research directions informed and inspired by this prior and current research. Bio: Sooyeon Lee is a postdoctoral research associate in the Golisano College of Computing and Information Sciences at Rochester Institute of Technology. She received her Ph.D., advised by Dr. John M. Carroll, in Information Sciences and Technology from the College of Information Sciences and Technology at The Pennsylvania State University, and she also conducted design research at Google and Uber. Her research is in the fields of Human-Computer Interaction and Human-AI Interaction with focus on accessibility. She designs, builds, and evaluates new systems and applications that address accessibility barriers. Her work investigates the diversity of users, explores and leverages emerging technologies, and adopts human-centered design and inclusive design approaches in an interdisciplinary research framework. She has multiple publications in top-tier human-computer interaction and computing accessibility journals and conferences, including ACM CHI, CSCW, ASSETS, and TACCESS, and she has received a Best Paper Award Nomination at ASSETS 2021. She has served on Associate Chair for the ACM CHI conference and will serve on Program Committee for ASSETS 2022.

Abstract: Pretraining vision encoders with self-supervision (SSL) leads to stronger representations that excel across diverse downstream tasks. One of the key factors enabling self-supervision is extracting multiple views of the same scene to formulate either: 1) View-invariant pretraining (DINO, SimCLR, iBOT), where the objective is predicting the same representation for different views of the scene; or 2) Cross-view pretraining (cross-view Masked Autoencoders), where the objective is predicting missing parts of one view using other views. For extracting multiple views, view-invariant methods rely on a combination of handcrafted augmentations (random cropping, color jittering, gaussian blur, etc.) of the same image, whereas cross-view pretraining methods rely on image cropping or video frames. In this work, we present methods to effectively incorporate synthetic views from diffusion models into SSL training.
For view-invariant pretraining, we introduce Gen-SIS, a method that leverages the ability of diffusion models to generate interpolated images through interpolation in conditioning space. We introduce a disentanglement pretext task: disentangling two source images from an interpolated synthetic image. This disentanglement task, in addition to vanilla single-source generative augmentation for view extraction, improves visual pretraining of various view-invariant methods (DINO, SimCLR, iBOT).
For cross-view pretraining, we introduce CDG-MAE, a novel cross-view masked autoencoder (MAE) based method that uses diverse synthetic views generated from static images via an image-conditioned diffusion model to learn dense correspondences. We present a quantitative method to evaluate the local and global consistency of the generated views to choose the right diffusion model for cross-view pretraining. These generated views exhibit substantial changes in pose and perspective, providing a rich training signal that overcomes the limitations of video (expensive) and crop-based (less variation) methods. CDG-MAE substantially narrows the gap to video-based MAE methods on video label propagation tasks while maintaining the data advantages of image-only MAEs.

Speaker: Varun Belagali

Location: NCS 120
Zoom: https://stonybrook.zoom.us/j/93647452432?pwd=hZaX7LXCAD8KPHWYE1Afw2sDI3owpv.1

Abstract: Much like other AI for Science domains, polymer design poses significant challenges. It requires grounding in empirical data and physical laws, precise handling of domain-specific structured representations, and compositional reasoning over multiple interacting constraints--all while working with limited data.

To address these limitations, we introduce PolyBench, a large-scale benchmark comprising over 125K polymer design and analysis tasks grounded in verified experimental and synthetic data. PolyBench includes tasks created from a wide range of data sources and presents diverse structural, property-driven, and synthesis-oriented reasoning problems. Tasks in PolyBench are organized from simple to complex analytical reasoning problems, enabling generalization tests and includes diagnostic probes to evaluate model capabilities. Additionally, to support effective domain alignment, we propose a knowledge-augmented reasoning distillation framework that enriches the dataset with structured chain-of-thought supervision derived from expert-informed reasoning strategies.

Small language models (7B-14B parameters) trained on PolyBench substantially outperform comparably sized baselines and, in many cases, exceed the performance of larger closed-source frontier models on polymer reasoning tasks, while also demonstrating improved transfer to external polymer benchmarks. Last, we conduct a diagnostic study that reveals a compositionality gap: despite strong performance on decomposed sub-questions, models struggle to integrate multiple interacting constraints and intermediate reasoning steps, highlighting fundamental limitations in current scientific language models.

Speaker: Dikshya Mohanty

Location: NCS 115/Online

Zoom: https://stonybrook.zoom.us/j/94746001760?pwd=BCAd8gu7cXLn3PXM6kkbh11V6r0Mr7.1
Meeting ID: 947 4600 1760 Passcode: 987917


Abstract: Human gaze behavior is a fundamental cue for understanding social intent, human-machine interaction, and cognitive processes. This dissertation addresses the challenges of gaze target estimation (GTE), also known as gaze following, by developing a holistic understanding of gaze across complex environments.
First, we improve GTE performance through Patch-level Distribution Prediction (PDP). Unlike traditional pixel-wise regression, PDP models gaze as a spatial distribution over patches, better accounting for annotation variance and regularizing the pixel-wise heatmap prediction through multi-scale modeling. Second, to mitigate the high cost of data labeling, we present GCDR, the first semi-supervised method for gaze following. By prompting large Visual Question Answering (VQA) models to generate initial Grad-CAM heatmaps and refining them via a diffusion model, GCDR achieves robust performance with minimal human annotation. Third, we expand the applicability of GTE to multi-camera environments. By introducing the Multi-View Gaze Target (MVGT) dataset, along with two novel frameworks for integrating information and predicting gaze targets across views, we explore a new direction that overcomes single-view limitations such as face occlusion and out-of-view targets. Finally, we propose OmniGF, a multi-person gaze following model built on Vision-Language Models (VLMs) that enriches gaze target localization with semantic and social reasoning. By leveraging the semantic capabilities of VLMs alongside structural innovations to ground the model with fine-grained cues for each individual, OmniGF achieves state-of-the-art performance across three gaze following tasks.
Collectively, by tackling the gaze following problem through the distinct yet complementary perspectives of probabilistic modeling, geometric reasoning, and multimodal learning, this dissertation builds a holistic understanding of human gaze, paving the way for more intuitive artificial intelligence systems in downstream applications.

Speaker: Qiaomu Miao

Location: NCS 220
University Libraries Present: AI as Author? New Considerations When Evaluating Sources.
In this workshop, librarian Christine Fena will review some ways AI is being integrated into published work within the worlds of news and scholarly publication, and discuss how this might impact how to evaluate and understand sources during the research process.
10/2 12:30-1:30 pm on Zoom.
Register via link: https://stonybrook.campuslabs.com/engage/event/10460202

Hear from Yvette Wohn, Ph.D., Inaugural Chair of the Department of Technology, AI and Society, for an engaging look at how artificial intelligence and emerging technologies are reshaping the way we live, work, interact, and connect, and get a sneak peek of how our upcoming programs will address this changing landscape.

Reception immediately following.