Hidden Biases. Ethical Issues in NLP, and What to Do about Them presented by Dirk Hovy of Bocconi University

ABSTRACT: Through language, we fundamentally express who we are as humans. This property makes text a fantastic resource for research into the complexity of the human mind, from social sciences to humanities. However, it is exactly that property that also creates some ethical problems. Texts reflect the authors' biases, which get magnified by statistical models. This has unintended consequences for our analysis: If our data is not reflective of the population as a whole, if we do not pay attention to the biases contained, we can easily draw the wrong conclusions, and create disadvantages for our users.

In this talk, I will discuss several types of biases that affect NLP models, their sources, and potential counter measures: (1) Bias stemming from data, i.e., selection bias (if our texts do not adequately reflect the population we want to study), label bias (if the labels we use are skewed) and semantic bias (the latent stereotypes encoded in embeddings); (2) Biases deriving from the models themselves, i.e., their tendency to amplify any imbalances that are present in the data; (3) Design bias, i.e., the biases arising from our (the researchers) decisions which topics to analyze, which data sets to use, and what to do with them. For each bias, I will provide examples and discuss the possible ramifications for a wide range of applications, and various ways to address and counteract these biases, ranging from simple labeling considerations to new types of models.

BIO: Dirk Hovey is an associate professor of Computer Science in the department of marketing at Bocconi University. He received his PhD from the University of Southern California in Los Angeles, where he worked as a research assistant at the Information Sciences Institute. 

He works in Natural Language Processing (NLP), a subfield of artificial intelligence. His research focuses on computational social science. His interests include integrating sociolinguistic knowledge into NLP models, using large-scale statistics to model the interaction between people's socio-demographic profile and their language use, and ethics for data science and algorithmic fairness.

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, January 7, 2025, 12:00 pm -- CDS, Bldg. 725, Training Room

Speakers

Chuntian Cao, CDS AID - Neural Network Potential (NNP) for Battery Electrolytes

Yeonju Go, NPP Physics - Generative AI for High-Energy Nuclear Physics

Gilchan Park, CDS AID - Graph RAG: Indexing, Retrieval and Generation

Join ZoomGov Meeting: https://bnl.zoomgov.com/j/1615289117?pwd=Hqkbj9itxWrFnkhZ8rQXHPInO2gxdF.1

Meeting ID: 161 528 9117
Passcode: 991382

Learn how SBU staff are using AI.

Over the past year, we've explored the AI tools and services available at Stony Brook. Now it's time to hear from colleagues who are putting these tools to work in their everyday roles. Join Tamara Gregorian, Herb Schramm, and Allie Seal moderated by David Ecker as they share how they use AI in their position, the tasks it helps them accomplish and the lessons they've learned along the way. This session is an opportunity to see practical examples, ask questions, and gather ideas that you can apply in your own work.

Register here.

Designed for faculty, staff, presidents, provosts, academic leaders, student affairs professionals, IT specialists, librarians, researchers, administrators, institutional decision-makers, and other higher education stakeholders, the conference highlights practical strategies institutions can implement now while exploring longer-term governance, policy, and ethical considerations. Participants will leave with concrete tools, cross-institutional insights, and collaborative connections that support mission-aligned AI innovation.

Hosted by: AAC&U

Location: Atlanta, GA and Virtual

Register here.

In this workshop, we will take a historical journey through the ideas that eventually led to today's large language models, such as ChatGPT and Claude.

Modern LLMs can seem overwhelmingly complex, but they didn't appear all at once. Many of the basic ideas behind them have been developing for more than a century. By following this history step by step, we can better understand what language models do and how they evolved.

In part 1 of this workshop, we will begin with Andrey Markov's early work on sequences and probability in 1913 and Claude Shannon's influential work on information and language in 1948. From there, we will explore key ideas such as Markov chains, N-grams, and the next-word generation process using simple, accessible examples.

In part 2 of the workshop, which is not in this workshop, we will also look at how an artificial neuron can learn to predict the next word, an important step toward neural language models that later became the foundation of modern LLMs and we will see what perplexity is.

The workshop is designed for anyone who is curious about where today's AI came from and how a century of ideas about language, probability, and computation helped make modern language models possible. This workshop targets students of any level who are interested in language models or history. No background in computer science or artificial intelligence is required. The hardest part is converting counted numbers to probabilities.

Our historical journey will cover roughly the period from 1918 to 1970s.

Location: Melville Library, Central Reading Room, Lab A

Register for the event on SB Engaged.

2026 MIC Symposium on Multilingualism

Language, Sustainability, and AI


Friday, October 16, 9:00am - 5:00pm | Humanties 1006

Presented by: Center for Multilingual and Intercultural Communication (MIC)

Co-Sponsored by: College of Arts and Sciences (CAS), AI Innovation Institute (AI3), Office of Diversity, Inclusion, and Intercultural Initiatives (DI3), The Sir Run Run Shaw Endowment, and Department of Asian and Asian American Studies (AAAS)


Opening Remarks: David Wrobel (Dean, College of Arts and Sciences) and Lav Varshney (Director, AI Innovation Institute)

Keynote Speakers: Salikoko Mufwene, David Wei Dai, and Susan Brennan

  • 9:30 - 10:15am: Keynote #1 - Salikoko S. Mufwene
  • 10:15 - 10:45am: Faculty Response Panel #1 (chaired by S. N. Sridhar)
  • 11:00 - 11:45am: Keynote #2 - David Wei Dai
  • 11:45 - 12:15pm: Faculty Response Panel #2 (chaired by Yi Wang)
  • 2:00 - 2:45 pm: Keynote #3 - Susan E. Brennan
  • 2:45 - 3:15pm: Faculty Response Panel #3 (chaired by Jiwon Hwang)
  • 3:30 - 4:45pm: Synthesis and Open Discussion with all speakers, panelists, and participants (moderated by Shyam Sharma)

If you need special accomodation, please contact chikako.nakamura@stonybrook.edu


Abstract: Pre-trained diffusion and flow matching models have made visual generation remarkably powerful, enabling high-fidelity synthesis of images and videos from natural language prompts. However, their behavior is still largely dictated by the pre-training data distribution and likelihood objective, which do not directly encode downstream desiderata such as fine-grained semantic alignment, controllability, or realism. This gap motivates post-training: starting from a base generator and further optimizing it with additional supervision signals derived from human or reward model preferences.This work presents post-training for visual generative models through two complementary case studies. First, Hummingbird addresses the problem of fine-grained contextual alignment in image-text-to-image generation. We introduce a multimodal context evaluator that scores the consistency between rich contextual descriptions and generated images, capturing fine-grained alignment beyond global CLIP similarity. By directly backpropagating these differentiable rewards through the diffusion sampler, Hummingbird substantially improves semantic faithfulness while preserving high visual quality.
Second, PISCES tackles post-training for text-to-video generation, where alignment is inherently semantic-spatio-temporal. We show that naive VLM-based rewards suffer from distributional mismatch and token-level misalignment, leading to reward hacking and suboptimal optimization. PISCES introduces a bi-objective, Optimal Transport (OT)-aligned reward module: distributional OT using Neural Optimal Transport to align text and video embedding distributions, and discrete, partial OT over a spatio-temporal cost matrix to capture semantic alignment at the token level. These rewards are integrated into both direct backpropagation and GRPO-style optimization to post-train state-of-the-art text-to-video generators. Together, Hummingbird and PISCES provide a unified view of how carefully designed visual reward models, coupled with OT-based representation alignment, can reliably improve the downstream behavior of pre-trained image and video generators.

Speaker: Minh Quan Le

Location: NCS 220

Zoom: https://stonybrook.zoom.us/j/94798224254?pwd=CFraer25qnpORbJ14aAVHRwaSJOjJM.1
Abstract: Robot control has evolved from optimization-based controllers---precise but task-specific---through deep reinforcement learning's learned policies, to Vision-Language-Action (VLA) models that leverage pretrained vision-language backbones for language-conditioned manipulation across diverse tasks.
Despite their promise, VLAs exhibit a critical limitation: they function primarily as trajectory learners rather than skill learners. Recent evaluations reveal that VLAs often fail when faced with even minor variations in object initialization or environmental conditions, suggesting they memorize specific trajectories rather than acquiring generalizable manipulation skills. Attempts to address this through 3D spatial representations have shown limited success, indicating that the missing component may be more fundamental than geometric understanding alone.
This work argues that World Models (WMs)---internal representations that predict future states given actions---constitute the missing piece for robust VLA systems. We present one completed contribution and two ongoing investigations.
We developed a dual-layer world model for human-robot interaction that anticipates both physical scene evolution and latent human preferences for assistive tasks. Building on these foundations, we present ongoing work probing VLA internal representations to verify implicit world model existence, and propose a WM-VLA integration approach operating in the native visual domain through embedding prediction and image decoding.
Together, these contributions and investigations establish a foundation for WM-VLA systems, pointing toward robust, generalizable robot policies.
Speaker: Jason Qin
Location: NCS 220
The overall purpose of this seminar is to bring together people with interests in Computer Vision theory and techniques and to examine current research issues. This course will be appropriate for people who already took a Computer Vision graduate course or already had research experience in Computer Vision. To enroll in this course, you must either: (1) be in the PhD program or (2) receive permission from the instructors. Each seminar will consist of multiple short talks (around 15 minutes) by multiple students. Students can register for 1 credit for CSE656. Registered students must attend and present a minimum of 2 talks. Everyone else is welcome to attend. Fill in https://forms.gle/q6UG9ygauLp2a8Po8 to subscribe to our mailing list for further announcement.