4 citations · 15 across the 17 of their papers we have counts for
17 papers
Counterparty Modeling is Not Strategy: The Limits of LLM Negotiators
Romain Cosentino, Sarath Shekkizhar, Adam Earle +1
Negotiation requires more than inferring what the other side wants: it requires using that information to make advantageous offers and counteroffers over multiple turns. We study w…
Beyond the Assistant Turn: User Turn Generation as a Probe of Interaction Awareness in Language Models
Sarath Shekkizhar, Romain Cosentino, Adam Earle
Standard LLM benchmarks evaluate the assistant turn: the model generates a response to an input, a verifier scores correctness, and the analysis ends. This paradigm leaves unmeasur…
PLATE: Plasticity-Tunable Efficient Adapters for Geometry-Aware Continual Learning
Romain Cosentino
We develop a continual learning method for pretrained models that \emph{requires no access to old-task data}, addressing a practical barrier in foundation model adaptation where pr…
Echoing: Identity Failures when LLM Agents Talk to Each Other
Sarath Shekkizhar, Romain Cosentino, Adam Earle +1
As large language model (LLM) based agents interact autonomously with one another, a new class of failures emerges that cannot be predicted from single agent performance: behaviora…
Convergence dynamics of Agent-to-Agent Interactions with Misaligned objectives
Romain Cosentino, Sarath Shekkizhar, Adam Earle
We develop and analyze a theoretical framework for agent-to-agent interactions in a simplified in-context linear regression setting. In our model, each agent is instantiated as a s…
AGI Is Coming... Right After AI Learns to Play Wordle
Sarath Shekkizhar, Romain Cosentino
This paper investigates multimodal agents, in particular, OpenAI's Computer-User Agent (CUA), trained to control and complete tasks through a standard computer interface, similar t…
Reasoning in Large Language Models: A Geometric Perspective
Romain Cosentino, Sarath Shekkizhar
The advancement of large language models (LLMs) for real-world applications hinges critically on enhancing their reasoning capabilities. In this work, we explore the reasoning abil…
Characterizing Large Language Model Geometry Helps Solve Toxicity Detection and Generation
Randall Balestriero, Romain Cosentino, Sarath Shekkizhar
Large Language Models (LLMs) drive current AI breakthroughs despite very little being known about their internal representations. In this work, we propose to shed the light on LLMs…
The Geometry of Self-supervised Learning Models and its Impact on Transfer Learning
Romain Cosentino, Sarath Shekkizhar, Mahdi Soltanolkotabi +2
Self-supervised learning (SSL) has emerged as a desirable paradigm in computer vision due to the inability of supervised models to learn representations that can generalize in doma…
Toward a Geometrical Understanding of Self-supervised Contrastive Learning
Romain Cosentino, Anirvan Sengupta, Salman Avestimehr +4
Self-supervised learning (SSL) is currently one of the premier techniques to create data representations that are actionable for transfer learning in the absence of human annotatio…
Spatial Transformer K-Means
Romain Cosentino, Randall Balestriero, Yanis Bahroun +3
K-means defines one of the most employed centroid-based clustering algorithms with performances tied to the data's embedding. Intricate data embeddings have been designed to push $…
Interpretable Image Clustering via Diffeomorphism-Aware K-Means
Romain Cosentino, Randall Balestriero, Yanis Bahroun +3
We design an interpretable clustering algorithm aware of the nonlinear structure of image manifolds. Our approach leverages the interpretability of -means applied in the image s…
Sparse Multi-Family Deep Scattering Network
Romain Cosentino, Randall Balestriero
In this work, we propose the Sparse Multi-Family Deep Scattering Network (SMF-DSN), a novel architecture exploiting the interpretability of the Deep Scattering Network (DSN) and im…
Nonlinear Regression with a Convolutional Encoder-Decoder for Remote Monitoring of Surface Electrocardiograms
Anton Banta, Romain Cosentino, Mathews M John +4
We propose the Nonlinear Regression Convolutional Encoder-Decoder (NRCED), a novel framework for mapping a multivariate input to a multivariate output. In particular, we implement…
Deep Autoencoders: From Understanding to Generalization Guarantees
Romain Cosentino, Randall Balestriero, Richard Baraniuk +1
A big mystery in deep learning continues to be the ability of methods to generalize when the number of model parameters is larger than the number of training examples. In this work…
The Geometry of Deep Networks: Power Diagram Subdivision
Randall Balestriero, Romain Cosentino, Behnaam Aazhang +1
We study the geometry of deep (neural) networks (DNs) with piecewise affine and convex nonlinearities. The layers of such DNs have been shown to be {\em max-affine spline operators…
Overcomplete Frame Thresholding for Acoustic Scene Analysis
Romain Cosentino, Randall Balestriero, Richard Baraniuk +1
In this work, we derive a generic overcomplete frame thresholding scheme based on risk minimization. Overcomplete frames being favored for analysis tasks such as classification, re…