Language Model Decoding as Likelihood–Utility Alignment

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.

Now is an exciting time for the domain of Earth observation (EO), with a multitude of diverse sensors looking at the planet from satellites, airplanes or drones. The volume of imagery acquired is massive, and hold great potential for a variety of applicati ...

The ability to reason, plan and solve highly abstract problems is a hallmark of human intelligence. Recent advancements in artificial intelligence, propelled by deep neural networks, have revolutionized disciplines like computer vision and natural language ...

Cells are essential to understanding health and disease, yet traditional models fall short of modeling and simulating their function and behavior. Advances in AI and omics offer groundbreaking opportunities to create an AI virtual cell (AIVC), a multi-scal ...

The rapid progress of industrial development, urbanization, and traffic has caused air quality degradation that negatively affects human health and environmental sustainability, especially in developed countries. However, due to the limited number of senso ...

Since the advent of internet and mass communication, two public-key cryptographic algorithms have shared the monopoly of data encryption and authentication: Diffie-Hellman and RSA. However, in the last few years, progress made in quantum physics -- and mo ...

We introduce an algorithm to reconstruct a mesh from discrete samples of a shape's Signed Distance Function (SDF). A simple geometric reinterpretation of the SDF lets us formulate the problem through a point cloud, from which a surface can be extracted wit ...

Multiple tensor-times-matrix (Multi-TTM) is a key computation in algorithms for computing and operating with the Tucker tensor decomposition, which is frequently used in multidimensional data analysis. We establish communication lower bounds that determine ...

The recent advance of large language models (LLMs) demonstrates that these large-scale foundation models achieve remarkable capabilities across a wide range of language tasks and domains. The success of the statistical learning approach challenges our unde ...

Language models (LMs) have recently shown remarkable performance on reasoning tasks by explicitly generating intermediate inferences, e.g., chain-of-thought prompting. However, these intermediate inference steps may be in- appropriate deductions from the i ...

Non-convex constrained optimization problems have become a powerful framework for modeling a wide range of machine learning problems, with applications in k-means clustering, large- scale semidefinite programs (SDPs), and various other tasks. As the perfor ...

Minimax-fair machine learning minimizes the error for the worst-off group. However, empirical evidence suggests that when sophisticated models are trained with standard empirical risk minimization (ERM), they often have the same performance on the worst-of ...

Machine learning models trained with passive sensor data from mobile devices can be used to perform various inferences pertaining to activity recognition, context awareness, and health and well-being. Prior work has improved inference performance through t ...

In this thesis, we give new approximation algorithms for some NP-hard problems arising in resource allocation and network design. As a resource allocation problem, we study the Santa Claus problem (also known as the MaxMin Fair Allocation problem) in which ...

With the looming threat of large-scale quantum computers, a fair portion of recent cryptographic research has focused on examining cryptographic primitives from the perspective of a quantum adversary. Shor's 1994 result revealed that quantum computers can ...

Finding cycles in directed graphs enables important applications in various domains such as finance, biology, chemistry, and network science. However, as the size of graph datasets continues to grow, it becomes increasingly difficult to discover cycles wit ...

Over the past two decades, epidural electrical stimulation (EES) has transitioned from a treatment for chronic pain to a promising paradigm for ameliorating deficits associated with neurological disorders like spinal cord injury (SCI) and Parkinson's Disea ...

Cycles are one of the fundamental subgraph patterns and being able to enumerate them in graphs enables important applications in a wide variety of fields, including finance, biology, chemistry, and network science. However, to enable cycle enumeration in r ...

The balancing of mechanisms consists in distributing their moving masses, inertias, and elastic components in order to achieve key mechanical properties, such as the elimination of the shaking forces and moments exported onto their supporting structure or ...

Single Domain Generalization (SDG) tackles the problem of training a model on a single source domain so that it generalizes to any unseen target domain. While this has been well studied for image classification, the literature on SDG object detection remai ...

Post-quantum cryptography is a branch of cryptography which deals with cryptographic algorithms whose hardness assumptions are not based on problems known to be solvable by a quantum computer, such as the RSA problem, factoring or discrete logarithms. This ...

Language Model Decoding as Likelihood–Utility Alignment

Chat with Graph Search

Interact with Earth Observation Images using AI: Transparent Methods and Evaluation for Visual Question Answering

Infusing structured knowledge priors in neural models for sample-efficient symbolic reasoning

How to build the virtual cell with artificial intelligence: Priorities and opportunities

Self-supervised air quality estimation with graph neural network assistance and attention enhancement

Secure and Efficient Cryptographic Algorithms in a Quantum World

Reach For the Arcs: Reconstructing Surfaces from SDFs via Tangent Points

COMMUNICATION LOWER BOUNDS AND OPTIMAL ALGORITHMS FOR MULTIPLE TENSOR-TIMES-MATRIX COMPUTATION

Monotonicity Reasoning in the Age of Neural Foundation Models

REFINER: Reasoning Feedback on Intermediate Representations

Augmented Lagrangian Methods for Provable and Scalable Machine Learning

When do Minimax-fair Learning and Empirical Risk Minimization Coincide?

Keep Sensors in Check: Disentangling Country-Level Generalization Issues in Mobile Sensor-Based Models with Diversity Scores

Approximation Algorithms for Allocation and Network Design

Security in the Presence of Quantum Adversaries

Fast Parallel Algorithms for Enumeration of Simple, Temporal, and Hop-Constrained Cycles

Closed-Loop Control of Epidural Electrical Stimulation in Patients with Spinal Cord Injury and Parkinson's Disease

Fast Parallel Algorithms for Enumeration of Simple, Temporal, and Hop-constrained Cycles

Mechanism balancing taxonomy

CLIP the Gap: A Single Domain Generalization Approach for Object Detection

Attacks on some post-quantum cryptographic protocols: The case of the Legendre PRF and SIKE