Unagi.py 勉強会37枚目～【祝4周年】機械学習・データサイエンスLT会～ - connpass
@hrs_sano645
来週はUnagi.py4周年のLT会です。LT募集してます。機械学習や分析、可視化系なら何でも（Pythonでなくても）OKです！
t.co/Ov9lQDTfEq
Machine Learning: The Great Stagnation - Mark Saroufim
@EcsuOist
“Here are the projects that I believe represent a glimmer of hope against the Stagnation of Machine Learning“ t.co/4cj81BNI2R
Machine Learning: The Great Stagnation - Mark Saroufim
@Jge_Ryu
This article is F I R E. If you are a ML practitioner you should give it a read.
#AI #MachineLearning #DeepLearning
t.co/whWWFG5XWn
KaggleOpsを考える ~ MLflow + Colaboratory + Kaggle Notebook ~ | GMOインターネット 次世代シ...
@currypurin
めちゃくちゃ便利になった。
あとは、このすごい仕組みを参考にできれば完璧だな
t.co/pnnW4G7e5E
How Over 200K Individuals Upleveled their Data and AI Expertise with Databr...
@databricks
We’ve had over 200K enroll in the Databricks Certification and Training programs 🌟
So many of you are growing your expertise! To help you show off your new skills, we’re excited to introduce digital badges. Learn more👇 t.co/6ElL0KjSNc
Towards Practical Second Order Optimization for Deep Learning | OpenReview
@_arohan_
We have posted our response to the AC.
t.co/8TwBE7ugPz
Saturn Cloud | 100x Faster Data Science & Machine Learning
@kdnuggets
Snowflake and Saturn Cloud Partner To Bring 100x Faster #DataScience to Millions of Python Users
t.co/YsYxWw3FgA t.co/Opsc95cgnC
Google AI Blog: ToTTo: A Controlled Table-to-Text Generation Dataset
@Wataoka_Koki
GoogleがToTTo(Table To Text)というデータセットを提案.
Wikipediaのテーブルからそれを忠実に説明するテキストを生成するというタスク.
paper: t.co/OFNx4bcaRR [EMNLP2020]
repo: t.co/JBFHZOvsDT
blog: t.co/jVkVA3IIAA t.co/lx7JjJSvPp
totto · Datasets at Hugging Face
@abhi1thakur
And its already available on @huggingface's datasets library 🤗
Check it out here: t.co/EaLFXuKDvD t.co/6JV8eNLe2U t.co/BRGClqaC65
New AI research to help predict COVID-19 resource needs from a series of X-...
@mattmucklm
We published a paper investigating MoCo-based self-supervised pretraining to predict COVID patient deterioration based on sequences of X-ray images. t.co/SjW6cEkZ9k
New AI research to help predict COVID-19 resource needs from a series of X-...
@EdgeImpulse
Researchers from @FacebookAI and @nyulangone have created machine learning models that could help doctors predict how a COVID-19 patient’s condition may develop: t.co/CUHTBdrg8A t.co/EGKt1lrOxW
New AI research to help predict COVID-19 resource needs from a series of X-...
@ylecun
Facebook AI & NYU Langone Health are open-sourcing a system that can predict if a Covid patient will require intensive care, from chest x-rays, up to four days in the future.
Paper: t.co/zA1Ig3i6WA
Github: t.co/mSe18J1lLk
Blog post: t.co/5yl2KJ1QHp
New AI research to help predict COVID-19 resource needs from a series of X-...
@alexvoica
It can be challenging for medical experts to predict the course of COVID-19 in a patient.
Today, in partnership with @NYULangone, we are developing AI models that use chest X-rays to predict a patient’s condition so hospitals can plan ahead:
t.co/rLiudGTcNn
[2101.04909] COVID-19 Deterioration Prediction via Self-Supervised Represen...
@ylecun
Facebook AI & NYU Langone Health are open-sourcing a system that can predict if a Covid patient will require intensive care, from chest x-rays, up to four days in the future.
Paper: t.co/zA1Ig3i6WA
Github: t.co/mSe18J1lLk
Blog post: t.co/5yl2KJ1QHp
[2101.04909] COVID-19 Deterioration Prediction via Self-Supervised Represen...
@farahshamout
The results of our collaboration with @facebookai are now published online. Check it out👇🏻
arXiv paper: t.co/jwcMJujoje
The code is open-source and can be found here: t.co/sxfMLf4wd3
@nyulangone @NYUADResearch t.co/eP34KRfNqN
Akira’s ML News #Week3, 2021. Here are some of the papers and… | by Akihiro...
@AkiraTOSEI
I post weekly ML report blog.
An interesting topic is the following:
- Fast training of large-scale models by switching layers for each token.
- High performance generative models by reduce differences in frequency space.
t.co/oMK9WmHsYM
Buttcher情報検索本メモ - Google Docs
@kosuke_tsujino
7章(動的転置インデックス)終わった。
ドキュメントが動的に追加・削除される状況でのインデックスの維持方法。インデックスを分割して持つことを許すことでマージ操作を楽にする。削除はフラグ付けしておいてあとでまとめてGC。妥当。
t.co/rC7iSh9u1d
Audio-Visual Learning in NeurIPS2020 - Speaker Deck
@ymas0315
NeurIPS2020で発表のあったAudio-Visualの自己教師あり学習について何件かまとめました．本日ExaWizardsさん主催のNeurIPS2020 オンライン読み会で発表します．
t.co/9Jnw9HfAFu
The Joint Conference of the 59th Annual Meeting of the Association for Comp...
@aclmeeting
Ready to submit to ACL-IJCNLP 2021? You can go to: t.co/JFMq7Ejk5T The abstract due is Jan 25, 2021 and paper due Feb 1, 2021. For more details, check the abstract form: t.co/6Jpo4TDah3 and paper submission form: t.co/uwUBNtYmlK
AI-SaMD Trend (2015-2020)
@Tdys13
エムスリーで医療AI領域の事業開発をしてる @stockholm_sy さんのトレンドまとめ。自分のスライドとは違った切り口で事業を捉えてて、とても勉強になる！！！あざす！！
スライドはこちら→t.co/l7n4GNnbxo t.co/nuI6iErB7u
AI-SaMD Trend (2015-2020)
@stockholm_sy
趣味でAI医療機器（主にAI-CAD）の研究開発・薬事・ビジネス化のトレンドをまとめてみました。あくまで概観をつかむためのものですので、ご参考まで。
AI-SaMD Trend (2015-2020) t.co/vUpp4RS8DK @SlideShareより
Google Cloudではじめる実践データエンジニアリング入門[業務で使えるデータ基盤構築]の通販/下田倫大/寳野雄太 - 紙の本：honto本の通...
@megumi_takahira
こっ、これはぜったい読まなきゃ…！！
hontoで予約doneです！！
t.co/TuSXQ6ryTu t.co/hP8U18nxoO
Practical Real Time Recurrent Learning with a Sparse Approximation | OpenRe...
@erich_elsen
So our paper (t.co/FSwgtsKm64) that was _desk_ rejected from NeurIPS was accepted, basically without changes, as a spotlight to ICLR. 🤔 Had one of my best and most productive conversations with this set of reviewers -- big thanks!
12 Best Machine Learning Books For Beginners | Book Riot
@BookRiot
If you have “expand my tech progress” on your 2021 to-do list, this list is for you: t.co/71QktICEKF
12 Best Machine Learning Books For Beginners | Book Riot
@7wData
12 Best #machinelearning Books For Beginners
New Year, New You, New…Skill If you’re looking to expanding your work beyond basic programming, it might be a great to ..
t.co/tkUK3qyC3L t.co/ZXdGUlaaBt
12 Best Machine Learning Books For Beginners | Book Riot
@touya_huji
Beginner領域はBig techのfree trainingが充実し過ぎていて、書籍で価値を出すことの難易度は上がっていると思うなあ。それでも突き抜けた良書は出ていて、領域の成熟も感じる
12 Best Machine Learning Books For Beginners- t.co/Pfhym2FbJV
GitHub - Chizuchizu/brain2
@chizu_potato
fujifilmコンペのコードです 大まかなsolutionはリプに書いておきます
未知の化合物の物性を実験なしで予測するというコンペでした
タスク2（水分配係数の予測）: t.co/0EQ59N5oeY
タスク3（タンパク質の結合予測）: t.co/yMoFAm3bp4
Rao-Blackwellizing the Straight-Through Gumbel-Softmax Gradient Estimator |...
@hillbig
潜在離散変数を含むNNの勾配を求める際、前向き計算はそのままで後ろ向き計算をGumbel SoftMaxで緩和した上でstraight-through推定を使う手法がSOTAだが、ラオ・ブラックウェル化を適用してして推定の分散をさらに抑えられる。特にバッチサイズが小さく温度が低い領域で有効t.co/pkXI3AA5uZ
Rao-Blackwellizing the Straight-Through Gumbel-Softmax Gradient Estimator |...
@hillbig
For the gradient estimation of NN with discrete latent variables, they apply Rao-Blackwellization to straight-through Gumbel-softmax to further reduce the variance. Effective in particular at small batch sizes and low temperatures. t.co/pkXI3AA5uZ
Recurrent Independent Mechanisms | OpenReview
@hillbig
Recurrent Independent Mechanisms (RIMs)はダイナミック環境の汎化可能なモデル化に向けて、複数の独立に動くRIMが入力から注意機構で読み出し、活性化したRIMのみ処理を行い、RIM間は疎な注意機構を使ってコミュニケーションする。導入での背景説明や目標説明が興味深い t.co/HYTO40OVvK
Recurrent Independent Mechanisms | OpenReview
@hillbig
For modeling dynamic environments, Recurrent Independent Mechanisms (RIMs) use a set of modules that operate independently and use attentions to read from the inputs, activate RIMs, and communicate between RIMs. Realization of the sparse factor graph. t.co/HYTO40OVvK
GitHub - datastacktv/data-engineer-roadmap: Roadmap to becoming a data engi...
@gtnao
このロードマップのやつ毎年見るけどデータエンジニアは初めて見た t.co/gQUxeOyJoT
VinBigData Chest X-ray Abnormalities Detection | Kaggle
@tereka114
これ考えてなかったのですが、No findingの扱い次第ではそうなるな‥
t.co/z2NyuIJeX5
VinBigData Chest X-ray Abnormalities Detection | Kaggle
@ZFPhalanx
やらかしたって思ったが、私の方ではscoreは上がらなかった。謎だ。
t.co/AMMJaIJmfA
PythonでグフラのGifアニメーションをつくる -データ分析のための可視化-｜はやぶさの技術ノート
@Cpp_Learning
『ダイエット×プログラミング』の記事を書きました
・ Pythonでどんなことができるの？
・データ可視化のポイントは？
・簡単なGifアニメの作り方とは？
などに興味のある人にオススメの記事です
PythonでグフラのGifアニメーションをつくる -データ分析のための可視化- t.co/txbJWNw9lo
Joel Grus – Fizz Buzz in Tensorflow
@tkasasagi
interviewer: Welcome, can I get you coffee or anything? Do you need a break?
me: No,I've probably had too much coffee already!
interviewer: Great, great. And are you OK with writing code on the whiteboard?
me:It's the only way I code!
interviewer: ...
t.co/VNGWErsYl0
Joel Grus – Fizz Buzz in Tensorflow
@cifar10
見覚えがあったので、元ネタを探した。
Fizz Buzz in Tensorflow
t.co/uoKx0MqYUN t.co/jY7jC0Y31a