1 citations · 2 across the 5 of their papers we have counts for
7 papers
Differentiable Architecture Pruning for Transfer Learning
Nicolo Colombo, Yang Gao
We propose a new gradient-based approach for extracting sub-architectures from a given large model. Contrarily to existing pruning methods, which are unable to disentangle the netw…
Adapting by Pruning: A Case Study on BERT
Yang Gao, Nicolo Colombo, Wei Wang
Adapting pre-trained neural models to downstream tasks has become the standard practice for obtaining high-quality models. In this work, we propose a novel model adaptation paradig…
Disentangling Neural Architectures and Weights: A Case Study in Supervised Classification
Nicolo Colombo, Yang Gao
The history of deep learning has shown that human-designed problem-specific networks can greatly improve the classification performance of general neural models. In most practical…
Training conformal predictors
Nicolo Colombo, Vladimir Vovk
Efficiency criteria for conformal prediction, such as \emph{observed fuzziness} (i.e., the sum of p-values associated with false labels), are commonly used to \emph{evaluate} the p…
Multiple Metric Learning for Structured Data
Nicolo Colombo
We address the problem of merging graph and feature-space information while learning a metric from structured data. Existing algorithms tackle the problem in an asymmetric way, by…
Counterfactual Distribution Regression for Structured Inference
Nicolo Colombo, Ricardo Silva, Soong M Kang +1
We consider problems in which a system receives external \emph{perturbations} from time to time. For instance, the system can be a train network in which particular lines are repea…