archimedes-Artificial Intelligence, Data Science, Algorithms-greece

 
Artificial Intelligence
 
Data Science
 
Algorithms

[Archimedes NLP Group Invited Talk]Understanding the Trade-Offs Between Hallucinations and Mode Collapse in Language Generation

Dates
2025-01-15 14:30 - 16:00
Venue
Artemidos 1 - Amphitheater
TITLE: Understanding the Trade-Offs Between Hallucinations and Mode Collapse in Language Generation

SPEAKER: Grigoris Velegkas(Yale University, USA)

ABSTRACT: Specifying all desirable properties of a language model is challenging, but certain requirements seem essential. Given samples from an unknown language, the trained model should produce valid strings not seen in the training set, and be expressive enough to capture the language's full breadth. Otherwise, outputting invalid strings constitutes "hallucination," and failing to capture the full breadth leads to "mode collapse." Recent work by Kleinberg and Mullainathan [KM24], building on classical work on the closely related problem of language identification by Gold [Gol67] and Angluin [Ang79, 80], provides a concrete mathematical framework to study the problem of language generation. Kleinberg and Mullainathan showed that for all countable collections of languages, it is possible to create a language model that does not hallucinate but suffers from mode collapse. They asked whether this tension between validity and breadth is inherent for language generation.

In this talk, we define various notions of breadth for language generation, and completely characterize when generation with validity and breadth is possible under each of these notions. Our results answer the question of [KM24] and show that this tension between validity and breadth is indeed inherent for language generation. Moreover, we formalize the notion of stable generation, a natural requirement derived from Gold’s work [Gold67], and discuss when this type of generation is achievable. Finally, we discuss the implications of our results in the universal rates setting of Bousquet, Hanneke, Moran, van Handel, and Yehudayoff [BGMvY21]. The talk is based on joint works with Alkis Kalavasis and Anay Mehrotra.

References:  https://arxiv.org/abs/2411.09642 , https://arxiv.org/abs/2412.18530 

SHORT BIO: Grigoris Velegkas is a final-year PhD student in Computer Science at Yale University, working with Prof. Amin Karbasi. Before that, he studied Electrical and Computer Engineering at the National Technical University of Athens, where he worked with Prof. Dimitris Fotakis. His research lies at the intersection of machine learning and theoretical computer science, and focuses on three main directions: i) understanding generalization properties of ML algorithms, ii) exploring responsible use of ML systems and designing algorithms with provable replicability guarantees, and iii) understanding the interaction between ML algorithms and mechanisms. He was a research intern at Google Research in summer 2023 and summer 2024, and a student researcher from October 2023 to May 2024.


________________________________________________________________________________
Microsoft Teams Need help?
Meeting ID: 364 928 182 762
Passcode: w24fHy
________________________________________________________________________________

 
 

Vision

To position Greece as a leading player in AI and Data Science

image
image

Mission

To build an AI Excellence Hub in Greece where the international research community can connect, groundbreaking ideas can thrive, and the next generation of scientists emerges, shaping a brighter future for Greece and the world

 

Welcome to ARCHIMEDES, a vibrant research hub connecting the global AI and Data Science research community fostering groundbreaking research in Greece and beyond. Its dedicated core team, comprising lead researchers, affiliated researchers, Post-Docs, PhDs and interns, is committed to advancing basic and applied research in Artificial Intelligence and its supporting disciplines, including Algorithms, Statistics, Learning Theory, and Game Theory organized around 8 core research areas. By collaborating with Greek and Foreign Universities and Research Institutes, ARCHIMEDES disseminates its research findings fostering knowledge exchange and providing enriching opportunities for students. Leveraging AI to address real-world challenges, ARCHIMEDES promotes innovation within the Greek ecosystem and extends its societal impact. Established in January 2022, as a research unit of the Athena Research Center with support from the Committee Greece 2021, ARCHIMEDES is funded for its first four years by the EU Recovery and Resilience Facility (RRF).

 
 

NEWS

 
Two Best Paper Awards at BIBE 2025

Two Best Paper Awards at BIBE 2025

12 November 2025

Researchers from the Archimedes Unit of the Athena Research Center, Greece, together with researchers from the Computer Science Department of the University of Crete, Greece; Stelios M. Smirnakis, Associate Neurologist at Brigham and Women’s Hospital, USA and Associate Professor at Harvard Medical School, USA; and Maria Papadopouli, Professor of Computer Science at the University of Crete, Greece, Affiliated Researcher at the Institute of Computer Science, FORTH, Crete, Greece and Lead Researcher at Archimedes, Athena Research Center, Greece, received two Best Paper Awards at the 25th annual IEEE International Conference on Bioinformatics and Bioengineering (BIBE 2025), which took place on November 6-8, 2025 in Athens, Greece.

10 Papers Accepted at EMNLP 2025!

10 Papers Accepted at EMNLP 2025!

5 November 2025

The Conference on Empirical Methods in Natural Language Processing (EMNLP) is a major annual conference for researchers in natural language processing, machine learning, and artificial intelligence. It has been organized by the Association for Computational Linguistics (ACL) Special Interest Group on Data (SIGDAT) since 1996 and is celebrating its 30th anniversary this year.

Archimedes Seminar by Mark Girolami, Chief Scientist of the Alan Turing Institute, UK

Archimedes Seminar by Mark Girolami, Chief Scientist of the Alan Turing Institute, UK

27 October 2025

On Friday 7 November, 2025, from 1:00 pm to 2:00 pm, at the Archimedes Amphitheatre (1 Artemidos Street, 15125, Marousi, Archimedes, Athena Research Center, Greece), Mark Girolami, Sir Kirby Laing Professor of Civil Engineering within the Department of Engineering at the University of Cambridge, UK, where he also holds the Royal Academy of Engineering Research Chair in Data Centric Engineering, and Chief Scientist of the Alan Turing Institute, UK, will deliver an Archimedes Seminar on "Statistical Finite Element Methods."

Archimedes Talk by Alexandra Meliou on

Archimedes Talk by Alexandra Meliou on "Data Analysis and Manipulation through a Constrained Optimization Lens"

22 October 2025

On Tuesday 4 November, 2025, from 1:00 pm to 2:00 pm, at the Archimedes Amphitheatre (1 Artemidos Street, 15125, Marousi, Archimedes, Athena Research Center, Greece), Professor Alexandra Meliou, Professor at the College of Information and Computer Sciences at the University of Massachusetts Amherst, USA, and a visiting researcher at the Archimedes Research Unit, Athena Research Center, Greece, will deliver an Archimedes talk on "Data Analysis and Manipulation through a Constrained Optimization Lens."

Exploring the Frontiers of AI and Biomechanics at ICIST 2025

Exploring the Frontiers of AI and Biomechanics at ICIST 2025

20 October 2025

Archimedes, Athena Research Center, Greece is happy to see its work being featured at the 15th International Conference on Information Society and Technology (ICIST 2025), at Kaunas University of Technology, Lithuania, through an invited keynote speech by Dr. Serafeim Moustakidis, titled: “Bringing Biomechanics into the AI Era: A New Frontier for Gait and Joint Health.”

More news items here
 
 

The project “ARCHIMEDES Unit: Research in Artificial Intelligence, Data Science and Algorithms” with code OPS 5154714 is implemented by the National Recovery and Resilience Plan “Greece 2.0” and is funded by the European Union – NextGenerationEU.

greece2.0 eu_arch_logo_en

 

Stay connected! Subscribe to our mailing list by emailing sympa@lists.athenarc.gr
with the subject "subscribe archimedes-news Firstname LastName"
(replace with your details)