7 papers
Exact Unlearning via Quantized Sufficient Statistics
Ami Tavory, Shripad Gade, Tal Sarig +2
Exact unlearning requires a deployed predictor to match one rebuilt without the information named by a deletion request. Existing general-purpose exact methods localize retraining…
Bandits via Additive Quantized Representations
Ami Tavory, Noam Touitou, Tal Sarig +2
Contextual bandits require balancing nonlinear reward modeling with online efficiency. Tree ensembles and neural methods capture nonlinearities but require periodic retraining and…
A Graph Meta-Network for Learning on Kolmogorov-Arnold Networks
Guy Bar-Shalom, Ami Tavory, Itay Evron +3
Weight-space models learn directly from the parameters of neural networks, enabling tasks such as predicting their accuracy on new datasets. Naive methods -- like applying MLPs to…
Billion-Scale Graph Foundation Models
Maya Bechler-Speicher, Yoel Gottlieb, Andrey Isakov +5
Graph-structured data underpins many critical applications. While foundation models have transformed language and vision via large-scale pretraining and lightweight adaptation, ext…
Multicalibration Yields Better Matchings
Riccardo Colini Baldeschi, Simone Di Gregorio, Simone Fioravanti +9
Consider the problem of finding the best matching in a weighted graph where we only have access to predictions of the actual stochastic weights, based on an underlying context. If…
Measuring multi-calibration
Ido Guy, Daniel Haimovich, Fridolin Linder +4
A suitable scalar metric can help measure multi-calibration, defined as follows. When the expected values of observed responses are equal to corresponding predicted probabilities,…
Active learning with biased non-response to label requests
Thomas Robinson, Niek Tax, Richard Mudd +1
Active learning can improve the efficiency of training prediction models by identifying the most informative new labels to acquire. However, non-response to label requests can impa…