邁向機器學習模型可評估性理論
Toward a Theory of Evaluability
Kunal Talwar 提出一套不要求測試集與訓練過程隔離的模型可評估性形式化框架,研究測試樣本可能已用於訓練時,能否準確評估黑盒模型。在只檢視模型對測試集輸出的設定下,準確評估可行當且僅當數據分佈的支持集較小,研究亦給出樣本複雜度的緊確上下界。少量額外輸入查詢可提升評估者對自然分佈及概念類別的評估能力,並與程式測試中的 self-correctors 有關。
AuthorsKunal Talwar
Machine learning models, particularly large foundation models, are increasingly evaluated along multiple axes that are relevant for their usage. In addition to accuracy measures appropriate for the specific task of interest, measures capturing safety, calibration and fairness properties are often of interest. The classical approach to these problems assumes that the test set has been held out from the training process, an assumption that is increasingly harder to justify for large foundation models. In this work, we formalize a notion of evaluability without this assumption, and ask if it is feasible to evaluate black-box models in settings where our test samples may have been used during training. We study this problem under different settings, capturing different levels of access to the model. We first study the natural setting, where the evaluator only looks at the output of the model on the test set. We fully characterize the class of distributions where accurate evaluation is possible, and show tight upper and lower bounds on the sample complexity of evaluation. Our results show that black-box evaluation in this setup is feasible if and only if the distribution is close to being small support. We then consider evaluation algorithms that can query the model on additional inputs and demonstrate connections to self-correctors studied in program testing. Using these, we show that a small amount of query access can increase the power of the evaluator for natural distributions and concept classes.
Related readings and updates.
Speech foundation models have recently achieved remarkable capabilities across a wide range of tasks. However, their evaluation remains disjointed across tasks and model types. Different models excel at distinct aspects of speech processing and thus require different evaluation protocols. This paper proposes a unified taxonomy that addresses the question: Which evaluation is appropriate for which model? The taxonomy defines three orthogonal axes:…
At the 2024 Worldwide Developers Conference, we introduced Apple Intelligence, a personal intelligence system integrated deeply into iOS 18, iPadOS 18, and macOS Sequoia.
Apple Intelligence is comprised of multiple highly-capable generative models that are specialized for our users’ everyday tasks, and can adapt on the fly for their current activity. The foundation models built into Apple Intelligence have been fine-tuned for user experiences such as writing and refining text, prioritizing and summarizing notifications, creating playful images for conversations with family and friends, and taking in-app actions to simplify interactions across apps.
來源:Apple Machine Learning Research · machinelearning.apple.com