Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
The conventional recipe for maximizing model accuracy is to (1) train multiple models with various hyperparameters and (2) pick the individual model which performs best on a held-out validation set, discarding the remainder. In this paper, we revisit the second step of this procedure in the context of fine-tuning large pre-trained models, where fine-tuned models often appear to lie in a single low
Deep Neural Networks and Tabular Data: A Survey
Heterogeneous tabular data are the most commonly used form of data and are essential for numerous critical and computationally demanding applications. On homogeneous data sets, deep neural networks have repeatedly shown excellent performance and have therefore been widely adopted. However, their adaptation to tabular data for inference or data generation tasks remains challenging. To facilitate fu
Monitoring feature attributions: How Google saved one of the largest ML services in trouble | Google Cloud Blog
Monitoring feature attributions: How Google saved one of the largest ML services in trouble An emergency in the largest MLOps at GoogleClaudiu Gruia is a software engineer at Google who works on machine learning (ML) models that recommend content to billions of users daily. In Oct 2019, Claudiu was notified by an alert from a monitoring service. A specific model feature (let us call this feature F
Mandoline: Model Evaluation under Distribution Shift
Machine learning models are often deployed in different settings than they were trained and validated on, posing a challenge to practitioners who wish to predict how well the deployed model will perform on a target distribution. If an unlabeled sample from the target distribution is available, along with a labeled sample from a possibly different source distribution, standard approaches such as im
AI Paper Recommendations from Experts
See part two with new experts here. After the 'top AI books' reading list was so well received, we reached out to some of our community to find out which papers they believe everyone should have read! All of the below papers are free to access and cover a range of topics from Hypergradients to modeling yield response for CNNs. Each expert also included a reason as to why the paper was picked as we
機械学習アプリケーションにおけるテストについて - Re:ゼロから始めるML生活
機械学習系の話題が多い昨今ですが、実際触ってみると期待した精度・結果が出ないなんてことはよくあることではないでしょうか。 機械学習特有の性質として、データ自体がモデルを変化させ、結果として業務に影響を与えたりします。 仮に、機械学習屋さんが精度が出るモデルを構築したと言っても、それを導入するときに、システム全体での品質の維持に苦労したりします。 ということで、不確実性の大きい機械学習系開発についての、設計・テスト戦略でどうやってリスクを低減していけるかが一つカギになってくると思い、方法論について勉強しましたので、そのメモです。 非常に参考にしたのはこちら。 arxiv.org テストそのもののテクニックなどは、一般的なテスト駆動開発に関する書籍を合わせてをご参考ください。 テスト駆動開発 作者:Kent Beck発売日: 2017/10/14メディア: 単行本（ソフトカバー） テスト駆動P
Machine Learning Casual Talks #10でMackerelのロール内異常検知について発表しました - yasuhisa's blog
メルカリさんのオフィスで開かれたMachine Learning Casual Talks (MLCT) #10に「教師なし学習によるMackerelの異常検知機能について 〜設計/運用/評価の観点から〜」というタイトルで登壇してきました。 MLCTは機械学習をサービスで運用していく知見を共有する勉強会です。YouTube等で動画配信を積極的にしてくださっていて、はてなの京都オフィスでも鑑賞会と称してランチタイムに同僚と発表を見させてもらっていました。普段から勉強させてもあっていた勉強会に、登壇という形でちょっとはお返しできているとうれしいです。登壇させて頂き、ありがとうございました! 私の発表資料はこちらです。スライド46枚ありますが、発表は15分だったので本番はこれの短縮バージョンで発表させてもらいました。 教師なし学習によるMackerelの異常検知機能について 〜設計/運用/評価の
GitHub - py-why/EconML: ALICE (Automated Learning and Intelligence for Causation and Economics) is a Microsoft Research project aimed at applying Artificial Intelligence concepts to economic decision making. One of its goals is to build a toolkit that com
EconML is a Python package for estimating heterogeneous treatment effects from observational data via machine learning. This package was designed and built as part of the ALICE project at Microsoft Research with the goal to combine state-of-the-art machine learning techniques with econometrics to bring automation to complex causal inference problems. The promise of EconML: Implement recent techniq
機械学習モデルの判断根拠の説明
第20回ステアラボ人工知能セミナー https://stair.connpass.com/event/109983/ 【講演動画】 https://youtu.be/Fgza_C6KphU 【講演タイトル】 機械学習モデルの判断根拠の説明 【講演概要】 本講演では、機械学習モデルの判断根拠を提示するための説明法について紹介する。高精度な認識・識別が可能な機械学習モデルは一般に非常に複雑な構造をしており、どのような基準で判断が下されているかを人間が窺い知ることは困難である。このようなモデルのブラックボックス性を解消するために、近年様々なモデルの説明法が研究・提案されてきている。本講演の前半ではまず近年の代表的な研究について紹介する。後半では、発表者の最近の研究として「ランダムフォレストの簡略化」と「モデル列挙」について紹介する。Read less
What are the measure for accuracy of multilabel data?
Consider a scenario where you are provided with KnownLabel Matrix and PredictedLabel matrix. I would like to measure the goodness of the PredictedLabel matrix against the KnownLabel Matrix. But the challenge here is that KnownLabel Matrix have few rows only one 1 and other few rows have many 1's (those instance are multi labeled). An example of KnownLabel Matrix is given below. A =[1 0 0 0 0 1 0 0
Companion webpage to the book "Mathematics for Machine Learning"
Companion webpage to the book "Mathematics for Machine Learning". Copyright 2020 by Marc Peter Deisenroth, A. Aldo Faisal, and Cheng Soon Ong. Published by Cambridge University Press. View the Project on GitHub View On GitHub Please link to this site using https://mml-book.com. Twitter: @mpd37, @AnalogAldo, @ChengSoonOng. We wrote a book on Mathematics for Machine Learning that motivates people to
Patterns for Research in Machine Learning | Ali Eslami
18 Jul 2012 Here I list a handful of code patterns that I wish I was more aware of when I started my PhD. Each on its own may seem pointless, but collectively they go a long way towards making the typical research workflow more efficient. And an efficient workflow makes it just that little bit easier to ask the research questions that matter. My guess is that these patterns will not only be useful
A Survey of Model Compression and Acceleration for Deep Neural Networks
Deep neural networks (DNNs) have recently achieved great success in many visual recognition tasks. However, existing deep neural network models are computationally expensive and memory intensive, hindering their deployment in devices with low memory resources or in applications with strict latency requirements. Therefore, a natural thought is to perform model compression and acceleration in deep n
リリース、障害情報などのサービスのお知らせ
最新の人気エントリーの配信
j次のブックマーク
k前のブックマーク
lあとで読む
eコメント一覧を開く
oページを開く
Announcing cleanlab: a Python Package for ML and Deep Learning on Datasets with Label Errors
arXiv:2011.09926v3 [cs.LG] 19 May 2022
Facebookにログイン
https://mml-book.github.io/book/mml-book.pdf
Alpacaの市場予測システムを支える技術 / Alpaca Market forecast System
AAAS
http://arxiv.org/pdf/1802.03426