paper-with-me

홈 › Papers

GULP: a prediction-based metric between representations

2022-10-12 · Enric Boix-Adsera, Hannah Lawrence, George Stepaniants, Philippe Rigollet

Comparing the representations learned by different neural networks has recently emerged as a key tool to understand various architectures and ultimately optimize them. In this work, we introduce GULP, a family of distance measures between representations that is explicitly motivated by downstream predictive tasks. By construction, GULP provides uniform control over the difference in prediction performance between two representations, with respect to regularized linear prediction tasks. Moreover, it satisfies several desirable structural properties, such as the triangle inequality and invariance under orthogonal transformations, and thus lends itself to data embedding and visualization. We extensively evaluate GULP relative to other methods, and demonstrate that it correctly differentiates between architecture families, converges over the course of training, and captures generalization performance on downstream linear tasks.

📄 PDF Abstract BibTeX arXiv:2210.06545

Code (1)

sgstepaniants/gulp 공식 구현 pytorch

Tasks

Prediction

Similar Papers 제목 키워드 기반

Uniform Kernel Prober

2025-02-11 · Soumya Mukherjee, Bharath K. Sriperumbudur

The ability to identify useful features or representations of the input data based on training data that achieves low prediction error on test data across multiple prediction tasks is considered the key to multitask lear…

Predictionregression

A Latent Space Metric for Enhancing Prediction Confidence in Earth Observation Data

2024-01-30 · Ioannis Pitsiorlas, Argyro Tsantalidou, George Arvanitakis, Marios Kountouris 외

This study presents a new approach for estimating confidence in machine learning model predictions, specifically in regression tasks utilizing Earth Observation (EO) data, with a particular focus on mosquito abundance (M…

Earth Observation

GULP: Solar-Powered Smart Garbage Segregation Bins with SMS Notification and Machine Learning Image Processing

2023-04-25 · Jerome B. Sigongan, Hamer P. Sinodlay, Shahida Xerxy P. Cuizon, Joanna S. Redondo 외

This study intends to build a smartbin that segregates solid waste into its respective bins. To make the waste management process more interesting for the end-users; to notify the utility staff when the smart bin needs t…

Management

Constrained belief updates explain geometric structures in transformer representations

2025-02-04 · Mateusz Piotrowski, Paul M. Riechers, Daniel Filan, Adam S. Shai

What computational structures emerge in transformers trained on next-token prediction? In this work, we provide evidence that transformers implement constrained Bayesian belief updating -- a parallelized version of parti…

Bayesian InferencePrediction

The Geometry of Tokens in Internal Representations of Large Language Models

2025-01-17 · Karthik Viswanathan, Yuri Gardinazzi, Giada Panerai, Alberto Cazzaniga 외

We investigate the relationship between the geometry of token embeddings and their role in the next token prediction within transformer models. An important aspect of this connection uses the notion of empirical measure,…