2020’s Top AI & Machine Learning Research Papers
TOPBOTS The Best of Applied Artificial Intelligence, Machine Learning, Automation, Bots, Chatbots Despite the challenges of 2020, the AI research community produced a number of meaningful technical breakthroughs. GPT-3 by OpenAI may be the most famous, but there are definitely many other research papers worth your attention. For example, teams from Google introduced a revolutionary chatbot, Meena,
CS231n Deep Learning for Computer Vision
(this page is currently in draft form) Visualizing what ConvNets learn Several approaches for understanding and visualizing Convolutional Networks have been developed in the literature, partly as a response the common criticism that the learned features in a Neural Network are not interpretable. In this section we briefly survey some of these approaches and related work. Visualizing the activation
ICLR2020ã®ç•°å¸¸æ¤œçŸ¥è«–æ–‡ã®ç´¹ä»‹ (2019/11/23)
[DL輪èªä¼š]data2vec: A General Framework for Self-supervised Learning in Speech,...
Image GPT
We find that, just as a large transformer model trained on language can generate coherent text, the same exact model trained on pixel sequences can generate coherent image completions and samples. By establishing a correlation between sample quality and image classification accuracy, we show that our best generative model also contains features competitive with top convolutional nets in the unsupe
ONNC
ONNC (Open Neural Network Compiler) is a retargetable compilation framework designed specifically for proprietary deep learning accelerators. Its software architecture expedites porting ONNC to any Deep Learning Accelerator (DLA) design that supports ONNX (Open Neural Network Exchange) operators. ONNC guarantees executability across every DLA by means of transforming ONNX models into DLA-specific
Searching for Activation Functions
The choice of activation functions in deep networks has a significant effect on the training dynamics and task performance. Currently, the most successful and widely-used activation function is the Rectified Linear Unit (ReLU). Although various hand-designed alternatives to ReLU have been proposed, none have managed to replace it due to inconsistent gains. In this work, we propose to leverage auto
Tiny ImageNet Visual Recognition Challenge
Welcome to the tiny ImageNet evaluation server. Tiny ImageNet Challenge is the default course project for Stanford CS231N. It runs similar to the ImageNet challenge (ILSVRC). The goal of the challenge is for you to do as well as possible on the Image Classification problem. You will submit your final predictions on a test set to this evaluation server and we will maintain a class leaderboard. Tiny
パラメータ数10å„„ï¼æœ€æ–°ã®å·¨å¤§ç”»åƒèªè˜ãƒ¢ãƒ‡ãƒ«ã€ŒBiTã€çˆ†èª• & 解説 - Qiita
Deleted articles cannot be recovered. Draft of this article would be also deleted. Are you sure you want to delete this article? オミータã§ã™ã€‚ツイッターã§äººå·¥çŸ¥èƒ½ã®ã“ã¨ã‚„他媒体ã§æ›¸ã„ã¦ã„る記事ãªã© を紹介ã—ã¦ã„ã¾ã™ã®ã§ã€äººå·¥çŸ¥èƒ½ã®ã“ã¨ã‚’ã‚‚ã£ã¨çŸ¥ã‚ŠãŸã„æ–¹ãªã©ã¯æ°—軽ã«@omiita_atiimoをフォãƒãƒ¼ã—ã¦ãã ã•ã„ï¼ ãƒ‘ãƒ©ãƒ¡ãƒ¼ã‚¿æ•°10å„„ï¼æœ€æ–°ã®å·¨å¤§ç”»åƒèªè˜ãƒ¢ãƒ‡ãƒ«ã€ŒBiTã€çˆ†èª• & 解説 2019å¹´12月24æ—¥ã®ã‚¯ãƒªã‚¹ãƒžã‚¹ã‚¤ãƒ–ã«arxiv上ã§Google Brainã‹ã‚‰æ–°ãŸãªç”»åƒèªè˜ãƒ¢ãƒ‡ãƒ«ãŒç™ºè¡¨ã•ã‚Œã¾ã—ãŸã€‚ãã®åã‚‚ BiT(=Big Transfer)。ãã®æ€§èƒ½ã¯2019å¹´ã«GoogleãŒå‡ºã—ãŸEfficientNet(拙著解説記事)を様々ãªãƒ‡ãƒ¼ã‚¿ã‚»ãƒƒãƒˆã§è¶…ãˆã‚‹ã»ã©
深層å¦ç¿’ã®æ¥µå°é‡å化ã¨ã„ã†æ–°ãŸãªã‚‹å¸Œæœ›
(スターウォーズã®ãƒ†ãƒ¼ãƒžã‚’ã‹ã‘ã€ã‚†ã£ãã‚Šã¨ã‚¹ã‚¯ãƒãƒ¼ãƒ«ã—ãªãŒã‚‰ãŠèªã¿ãã ã•ã„) ã“ã®è¨˜äº‹ã¯ LeapMind Advent Calendar 2019 25日目ã®è¨˜äº‹ã§ã™ã€‚ ã‚‚ã¯ã‚„å¹´1回ã®æ›´æ–°ã™ã‚‰æ€ªã—ããªã£ã¦ããŸã“ã®ãƒ–ãƒã‚°ã§ã™ãŒã€ä»Šå¹´ã‚‚ãªã‚“ã¨ã‹æ›´æ–°ã§ãã¾ã—ãŸã€‚æ¥å¹´ã¯ã©ã†ãªã‚‹ã‹ãªã€‚ 概è¦LeapMindã¯æ¥µå°é‡å化ã¨ã„ã†æŠ€è¡“ã«ã‹ãªã‚Šæ³¨åŠ›ã—ã¦ã„ã¾ã™ãŒã€ã©ã†ã„ã†ã“ã¨ã‚’ã‚„ã£ã¦ã„ã‚‹ã®ã‹ã€ãªãœãã®ã‚ˆã†ãªã“ã¨ã‚’ã‚„ã£ã¦ã„ã‚‹ã®ã‹ã€CTO個人ã¨ã—ã¦ã®è¦–点ã‹ã‚‰è§£èª¬ã—ã¾ã™ã€‚ ã¯ã˜ã‚ã«LeapMindã¨ã„ã†ä¼šç¤¾ã®ç©¶æ¥µçš„ãªç›®æ¨™ã‚’端的ã«è¡¨ç¾ã™ã‚‹ã¨ã€æ©Ÿæ¢°ã‚’今よりもインテリジェントã«ã™ã‚‹ã“ã¨ã€ç‰¹ã«ã€ãƒ‡ãƒ¼ã‚¿ã‚»ãƒ³ã‚¿ãƒ¼ã«ã‚るよã†ãªè¨ˆç®—æ©Ÿã§ã¯ãªãã€èº«è¿‘ãªã‚¨ãƒƒã‚¸ãƒ‡ãƒã‚¤ã‚¹ã‚’インテリジェントã«ã™ã‚‹ã“ã¨ã§ã™ã€‚æ£ç¢ºãªãƒ‹ãƒ¥ã‚¢ãƒ³ã‚¹ã«ã¤ã„ã¦çŸ¥ã‚ŠãŸã„æ–¹ã¯ã€ç¾åœ¨çµ¶è³›ç¤¾å“¡å‹Ÿé›†ä¸ãªã®ã§ã€ãœã²ä¼šç¤¾è¦‹å¦ã«ã„らã£ã—ゃã£ã¦ãã ã•ã„。 ã•ã¦ã€ã“ã®ã‚ˆã†ãªå´‡é«˜ãªãƒŸãƒƒã‚·ãƒ§
CVPR 2019 é€Ÿå ±
CVPR 2019 ( http://cvpr2019.thecvf.com/ )ã®å‚åŠ é€Ÿå ±ã‚’æ›¸ãã¾ã—ãŸã€‚ ã“ã®è³‡æ–™ã«ã¯ä¸‹è¨˜ã®é …ç›®ãŒå«ã¾ã‚Œã¦ã„ã¾ã™ã€‚ ・DNNã®å‹•å‘・CVã®ãƒˆãƒ¬ãƒ³ãƒ‰ï¼ˆDNNã®æµã‚Œã¨CVã«ãŠã‘るトレンドã®å¤‰é·ï¼‰ ・CVPR 2019ã§ã®å‹•å‘や気付ã ・最近ã®ãƒˆãƒ¬ãƒ³ãƒ‰ã‚’創ã£ã¦ã„ã‚‹/引用ã•ã‚Œãã†ãªè«–æ–‡ ・今後ã®æ–¹é‡ （・CVPR 2019ã®è«–æ–‡ã¾ã¨ã‚ã¯ã“ã¡ã‚‰ã«ã‚ã‚Šã¾ã™ Link: http://xpaperchallenge.org/cv/survey/cvpr2019_summaries/ ） cvpaper.challengeã¯ã‚³ãƒ³ãƒ”ュータビジョン分野ã®ä»Šã‚’æ˜ ã—ã€ãƒˆãƒ¬ãƒ³ãƒ‰ã‚’創り出ã™æŒ‘戦ã§ã™ã€‚論文サマリ・アイディア考案・è°è«–・実装・論文投稿ã«å–り組ã¿ã€å‡¡ã‚†ã‚‹çŸ¥è˜ã‚’共有ã—ã¦ã„ã¾ã™ã€‚ http://xpaperchallenge.org/cv/
深層å¦ç¿’ã®ãƒ¢ãƒ‡ãƒ«åœ§ç¸®ãƒ»é«˜é€ŸåŒ–ã«é–¢ã™ã‚‹è«–æ–‡80本ノック - ãã†ã ã
ã“ã‚“ã«ã¡ã¯.無事ã«3年生ã«é€²ç´šã§ãã¦ãƒ›ãƒƒã¨ã—ã¦ã„ã¾ã™. 昨年度ã‹ã‚‰æ·±å±¤å¦ç¿’ã®ãƒ¢ãƒ‡ãƒ«æ”¹å¤‰ã«ã‚ˆã‚‹é«˜é€ŸåŒ–手法ã«ã¤ã„ã¦èª¿ã¹ã¦ã„ã¾ã—ãŸ.ãã‚“ãªã‚ã‘ã§æœ¬ç¨¿ã§ã¯æ·±å±¤å¦ç¿’モデルã®ãƒ‘ラメータ容é‡ã®å‰Šæ¸›ã‚„高速化を目的ã¨ã—ãŸæ‰‹æ³•ã«é–¢ã™ã‚‹è«–文を80本紹介ã—ã¾ã™. 一部, ãƒã‚¤ãƒ‘ーパラメータ探索(Neural Architecture Searchç‰)ã®è«–文を「高速ãªãƒ¢ãƒ‡ãƒ«ã®æŽ¢ç´¢ã«ä½¿ãˆãªã„ã‹ãªã€œã€ã¨ã„ã†ãã‚‚ã¡ã§å«ã‚ã¾ã—ãŸ. ã¾ãŸ, 蒸留ã«é–¢ã™ã‚‹è«–æ–‡ã¯å¿…ãšã—も高速化を狙ã£ãŸã‚‚ã®ã ã‘ã§ã¯ãªã, 蒸留ãã®ã‚‚ã®ã®æ´¾ç”Ÿè«–æ–‡ã«ã¤ã„ã¦ã‚‚å«ã‚ã¦ã„ã¾ã™. å…¨ã¦ã®è«–æ–‡ã«ã¤ã„ã¦è©³è§£ã§ã¯ãªã雑ãªæ¦‚è¦ã®ã¿è¨˜è¼‰ã—ã¦ã„ã¾ã™. サーベイã™ã‚‹éš›ã®ã‚¬ã‚¤ãƒ‰ã«ã§ã‚‚使ã£ã¦ã„ãŸã ã‘ã‚Œã°å¹¸ã„ã§ã™. サクッã¨èªã‚“ã ã ã‘ã®è«–文も多々ã‚ã‚Šã¾ã™ã®ã§é–“é•ã„ç‰ã”ã–ã„ã¾ã—ãŸã‚‰ã”指摘ãã ã•ã„. ã§ã¯å‚ã‚Šã¾ã—ょã†. 高速ãªãƒ¢ãƒ‡ãƒ«ã‚¢ãƒ¼ã‚テクãƒãƒ£ より良ã„速度ã¨ç²¾åº¦ã®ãƒˆãƒ¬ãƒ¼ãƒ‰ã‚ªãƒ•ã‚’
畳ã¿è¾¼ã¿ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®æœ€æ–°ç ”ç©¶å‹•å‘ (〜2017) - Qiita
Deleted articles cannot be recovered. Draft of this article would be also deleted. Are you sure you want to delete this article? 昨年ã«å¼•ã続ãDeep Learningã‚„ã£ã¦ã„ã Advent Calendar 2017ã®25日目ã®æ»‘ã‚Šè¾¼ã¿è¨˜äº‹ã§ï¼Œç•³ã¿è¾¼ã¿ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®æ´å²ãŠã‚ˆã³æœ€æ–°ã®ç ”究動å‘ã«ã¤ã„ã¦ã®ã‚µãƒ¼ãƒ™ã‚¤ã§ã™ï¼Ž2017å¹´12月ã®PRMUç ”ç©¶ä¼šã«ã¦ç™ºè¡¨ã—ãŸï¼Œç•³ã¿è¾¼ã¿ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã«ã¤ã„ã¦ã®ã‚µãƒ¼ãƒ™ã‚¤ã‚’ベースã«è¿½è¨˜ã‚’è¡Œã£ãŸã‚‚ã®ã«ãªã‚Šã¾ã™ï¼Ž ã¯ã˜ã‚㫠畳ã¿è¾¼ã¿ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ (Convolutional Neural Networks; CNN) （以é™CNN）ã¯ï¼Œä¸»ã«ç”»åƒèªè˜ã«åˆ©ç”¨ã•ã‚Œã‚‹ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®ä¸€ç¨®ã§ã‚る．CNNã®åŽŸåž‹ã¯
ランã‚ング
ランã‚ング
メンテナンス
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
http://image-net.org/challenges/LSVRC/2012/supervision.pdf
物体検出フレームワークMMDetectionã§å¿«é©ãªé–‹ç™º
GitHub - sekwiatkowski/awesome-capsule-networks: A curated list of awesome resources related to capsule networks
[DL Hacks] Hessian based-analysis
DLAU: A Scalable Deep Learning Accelerator Unit on FPGA
A Survey of FPGA Based Deep Learning Accelerators: Challenges and Opportunities
Why GEMM is at the heart of deep learning
Tensor flow usergroup 2016 (公開版)
{{#tags}}- {{label}}
{{/tags}}