paper-with-me

Papers

Backward Compatibility During Data Updates by Weight Interpolation

2023-01-25 · Raphael Schumann, Elman Mansimov, Yi-An Lai, Nikolaos Pappas, Xibin Gao, Yi Zhang

Backward compatibility of model predictions is a desired property when updating a machine learning driven application. It allows to seamlessly improve the underlying model without introducing regression bugs. In classification tasks these bugs occur in the form of negative flips. This means an instance that was correctly classified by the old model is now classified incorrectly by the updated model. This has direct negative impact on the user experience of such systems e.g. a frequently used voice assistant query is suddenly misclassified. A common reason to update the model is when new training data becomes available and needs to be incorporated. Simply retraining the model with the updated data introduces the unwanted negative flips. We study the problem of regression during data updates and propose Backward Compatible Weight Interpolation (BCWI). This method interpolates between the weights of the old and new model and we show in extensive experiments that it reduces negative flips without sacrificing the improved accuracy of the new model. BCWI is straight forward to implement and does not increase inference cost. We also explore the use of importance weighting during interpolation and averaging the weights of multiple new models in order to further reduce negative flips.

📄 PDF Abstract BibTeX arXiv:2301.10546

Code (0)

등록된 구현이 없습니다.

Tasks

regression

Similar Papers 제목 키워드 기반

Multi-modal Knowledge Preserving Adapter for Embedding Backward Compatibility

2026-09-15 · Jaeseok Byun, Gukyeong Kwon, Han-Kai Hsu, Meher Gitika Karumuri 외 arxiv

Upgrading embedding models typically requires expensive database re-indexing, as new query embeddings are incompatible with existing database embeddings. While Backward Compatible Training (BCT) mitigates this by enforci…

Video Retrieval

An Empirical Analysis of Backward Compatibility in Machine Learning Systems

2020-08-11 · Megha Srivastava, Besmira Nushi, Ece Kamar, Shital Shah 외

In many applications of machine learning (ML), updates are performed with the goal of enhancing model performance. However, current practices for updating models rely solely on isolated, aggregate performance analyses, o…

BIG-bench Machine Learning

Backward-Compatible Aligned Representations via an Orthogonal Transformation Layer

2024-08-16 · Simone Ricci, Niccolò Biondi, Federico Pernici, Alberto del Bimbo

Visual retrieval systems face significant challenges when updating models with improved representations due to misalignment between the old and new representations. The costly and resource-intensive backfilling process i…

A Case for Backward Compatibility for Human-AI Teams

2019-06-04 · Gagan Bansal, Besmira Nushi, Ece Kamar, Dan Weld 외

AI systems are being deployed to support human decision making in high-stakes domains. In many cases, the human and AI form a team, in which the human makes decisions after reviewing the AI's inferences. A successful par…

Decision Making

Towards Backward-Compatible Representation Learning

2020-03-26 · CVPR 2020 6 · Yantao Shen, Yuanjun Xiong, Wei Xia, Stefano Soatto

We propose a way to learn visual features that are compatible with previously computed ones even when they have different dimensions and are learned via different neural network architectures and loss functions. Compatib…

Face RecognitionRepresentation Learning