state-of-the-art-result-for-machine-learning-problems

KO | EN

gitlite — search

#javascript #typescript #llm #android #machine-learning #nodejs #awesome #cli #vue #awesome-list #security #pytorch #mcp #openai

state-of-the-art-result-for-machine-learning-problems

(★ 8,894)

이 저장소는 모든 머신러닝 문제에 대해 최고 수준(SoTA)의 결과를 제공합니다. 저희는 이 저장소를 최신 상태로 유지하기 위해 최선을 다하고 있습니다. 만약 어떤 문제의 SoTA 결과가 최신이 아니거나 누락된 것을 발견하시면, 이슈를 제기해 주시거나 Google 폼을 제출해 주세요(연구 논문 이름, 데이터셋, 평가지표, 소스 코드 및 연도 정보 포함). 즉시 수정하겠습니다.

파일 탐색기

최종 버전 다운로드 (.zip)

LICENSE
README.md

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# 모든 Machine Learning 문제를 위한 최첨단(State-of-the-art) 결과
### LAST UPDATE: 2019년 2월 20일
### NEWS: NLP, Computer Vision, Reinforcement Learning 분야를 연구하는 협업자를 찾고 있습니다. 연구원은 아니지만 의욕이 있으신 분들도 연락해 주세요. 이메일: yxt.stoaml@gmail.com
이 저장소는 모든 머신러닝 문제에 대한 최고 수준(SoTA)의 결과를 제공합니다. 본 저장소를 최신 상태로 유지하기 위해 최선을 다하고 있습니다. 만약 어떤 문제의 SoTA 결과가 오래되었거나 누락된 것을 발견하셨다면, (연구 논문 이름, 데이터셋, 성능 지표(metric), 소스 코드, 연도 등의 정보를 포함하여) 이슈를 제기해 주세요. 즉시 수정하겠습니다.
Github을 잘 모르시는 경우, 이 [Google Form](https://docs.google.com/forms/d/e/1FAIpQLSe\_fFZVCeCVRGGgOQIpoQSXY7mZWynsx7g6WxZEVpO5vJioUA/viewform?embedded=true)을 제출하셔도 됩니다.
이 프로젝트는 모든 종류의 머신러닝 문제에 대한 최고 수준의 결과를 한곳에 모으려는 시도입니다. 저 혼자서는 할 수 없습니다. 모든 분들의 도움이 필요합니다. 데이터셋에 대한 SOTA 결과를 찾으셨다면 Google Form을 제출해 주시거나 이슈를 제기해 주세요. 이 내용을 Twitter, Facebook 및 기타 소셜 미디어에 공유해 주세요.
이 요약은 다음 범주로 분류됩니다:
- [Supervised Learning](https://github.com/RedditSota/state-of-the-art-result-for-machine-learning-problems#supervised-learning)
- [Speech](https://github.com/RedditSota/state-of-the-art-result-for-machine-learning-problems#speech)
- [Computer Vision](https://github.com/RedditSota/state-of-the-art-result-for-machine-learning-problems#computer-vision)
- [NLP](https://github.com/RedditSota/state-of-the-art-result-for-machine-learning-problems#nlp)
- [Semi-supervised Learning](https://github.com/RedditSota/state-of-the-art-result-for-machine-learning-problems#semi-supervised-learning)
- Computer Vision
- [Unsupervised Learning](https://github.com/RedditSota/state-of-the-art-result-for-machine-learning-problems#unsupervised-learning)
- Speech
- Computer Vision
- [NLP](https://github.com/RedditSota/state-of-the-art-result-for-machine-learning-problems/blob/master/README.md#nlp-1)
- [Transfer Learning](https://github.com/RedditSota/state-of-the-art-result-for-machine-learning-problems#transfer-learning)
- [Reinforcement Learning](https://github.com/RedditSota/state-of-the-art-result-for-machine-learning-problems#reinforcement-learning)
## Supervised Learning
### NLP
#### 1. Language Modelling

| Research Paper | Datasets | Metric | Source Code | Year |
| --- | --- | --- | --- | --- |
| [Language Models are Unsupervised Multitask Learners](https://d4mucfpksywv.cloudfront.net/better-language-models/language-models.pdf) | * PTB * WikiText-2 | * Perplexity: 35.76 * Perplexity: 18.34 | [Tensorflow](https://github.com/openai/gpt-2) | 2019 |
| [BREAKING THE SOFTMAX BOTTLENECK: A HIGH-RANK RNN LANGUAGE MODEL](https://arxiv.org/pdf/1711.03953.pdf) | * PTB * WikiText-2 | * Perplexity: 47.69 * Perplexity: 40.68 | [Pytorch](https://github.com/zihangdai/mos) | 2017 |
| [DYNAMIC EVALUATION OF NEURAL SEQUENCE MODELS](https://arxiv.org/pdf/1709.07432.pdf) | * PTB * WikiText-2 | * Perplexity: 51.1 * Perplexity: 44.3 | [Pytorch](https://github.com/benkrause/dynamic-evaluation) | 2017 |
| [Averaged Stochastic Gradient Descent   with Weight Dropped LSTM or QRNN](https://arxiv.org/pdf/1708.02182.pdf) | * PTB * WikiText-2 | * Perplexity: 52.8 * Perplexity: 52.0 | [Pytorch](https://github.com/salesforce/awd-lstm-lm) | 2017 |
| [FRATERNAL DROPOUT](https://arxiv.org/pdf/1711.00066.pdf) | * PTB * WikiText-2 | * Perplexity: 56.8 * Perplexity: 64.1 | [Pytorch](https://github.com/kondiz/fraternal-dropout) | 2017 |
| [Factorization tricks for LSTM networks](https://arxiv.org/pdf/1703.10722.pdf) | One Billion Word Benchmark | Perplexity: 23.36 | [Tensorflow](https://github.com/okuchaiev/f-lm) | 2017 |

#### 2. Machine Translation

| 연구 논문 | 데이터셋 | 성능 지표 | 소스 코드 | 연도 |
| --- | --- | --- | --- | --- |
| [Understanding Back-Translation at Scale](https://arxiv.org/pdf/1808.09381v2.pdf) | * WMT 2014 English-to-French * WMT 2014 English-to-German | * BLEU: 45.6 * BLEU: 35.0 | * [PyTorch](https://github.com/pytorch/fairseq) | 2018 |
| [WEIGHTED TRANSFORMER NETWORK FOR MACHINE TRANSLATION](https://arxiv.org/pdf/1711.02132.pdf) | * WMT 2014 English-to-French * WMT 2014 English-to-German | * BLEU: 41.4 * BLEU: 28.9 | * NOT FOUND | 2017 |
| [Attention Is All You Need](https://arxiv.org/abs/1706.03762) | * WMT 2014 English-to-French * WMT 2014 English-to-German | * BLEU: 41.0 * BLEU: 28.4 | * [PyTorch](https://github.com/jadore801120/attention-is-all-you-need-pytorch) * [Tensorflow](https://github.com/tensorflow/tensor2tensor) | 2017 |
| [NON-AUTOREGRESSIVE NEURAL MACHINE TRANSLATION](https://einstein.ai/static/images/pages/research/non-autoregressive-neural-mt.pdf) | * WMT16 Ro→En | * BLEU: 31.44 | * [PyTorch](https://github.com/salesforce/nonauto-nmt) | 2017 |
| [Improving Neural Machine Translation with Conditional Sequence Generative Adversarial Nets](https://arxiv.org/abs/1703.04887) | * NIST02 * NIST03 * NIST04 * NIST05 | - 38.74 - 36.01 - 37.54 - 33.76  * [NMTPY](https://github.com/ngohoanhkhoa/GAN-NMT) | 2017 | |

#### 3. 텍스트 분류 (Text Classification)

| 연구 논문 | 데이터셋 | 성능 지표 | 소스 코드 | 연도 |
| --- | --- | --- | --- | --- |
| [Learning Structured Text Representations](https://arxiv.org/abs/1705.09207) | Yelp | Accuracy: 68.6 | * [Tensorflow](https://github.com/nlpyang/structured) | 2017 |
| [Attentive Convolution](https://arxiv.org/pdf/1710.00519.pdf) | Yelp | Accuracy: 67.36 | * [Theano](https://github.com/yinwenpeng/Attentive_Convolution) | 2017 |

#### 4. 자연어 추론 (Natural Language Inference)
리더보드:
[Stanford Natural Language Inference (SNLI)](https://nlp.stanford.edu/projects/snli/)
[MultiNLI](https://www.kaggle.com/c/multinli-matched-open-evaluation/leaderboard)

| 연구 논문 | 데이터셋 | 성능 지표 | 소스 코드 | 연도 |
| --- | --- | --- | --- | --- |
| [NATURAL LANGUAGE INFERENCE OVER INTERACTION SPACE](https://arxiv.org/pdf/1709.04348.pdf) | Stanford Natural Language Inference (SNLI) | Accuracy: 88.9 | [Tensorflow](https://github.com/YichenGong/Densely-Interactive-Inference-Network) | 2017 |
| [BERT-LARGE (ensemble)](https://arxiv.org/pdf/1810.04805.pdf) | Multi-Genre Natural Language Inference (MNLI) | * Matched accuracy: 86.7 * Mismatched accuracy: 85.9 | * [Tensorflow](https://github.com/google-research/bert) * [PyTorch](https://github.com/huggingface/pytorch-pretrained-BERT) | 2018 |

#### 5. 질의 응답 (Question Answering)
리더보드
[SQuAD](https://rajpurkar.github.io/SQuAD-explorer/)

| 연구 논문 | 데이터셋 | 성능 지표 | 소스 코드 | 연도 |
| --- | --- | --- | --- | --- |
| [BERT-LARGE (ensemble)](https://arxiv.org/pdf/1810.04805.pdf) | The Stanford Question Answering Dataset | * Exact Match: 87.4 * F1: 93.2 | * [Tensorflow](https://github.com/google-research/bert) * [PyTorch](https://github.com/huggingface/pytorch-pretrained-BERT) | 2018 |

#### 6. 개체명 인식 (Named entity recognition)

| 연구 논문 | 데이터셋 | 성능 지표 | 소스 코드 | 연도 |
| --- | --- | --- | --- | --- |
| [Named Entity Recognition in Twitter using Images and Text](https://arxiv.org/pdf/1710.11027.pdf) | Ritter | * F-measure: 0.59 | NOT FOUND | 2017 |

#### 7. 추상적 요약 (Abstractive Summarization)
연구 논문 | 데이터셋 | 성능 지표 | 소스 코드 | 연도
------------ | ------------- | ------------ | ------------- | -------------
[Cutting-off redundant repeating generations for neural abstractive summarization](https://aclanthology.info/pdf/E/E17/E17-2047.pdf) |

* DUC-2004
* Gigaword

|

* DUC-2004

+ ROUGE-1: \*\*32.28\*\*
+ ROUGE-2: 10.54
+ ROUGE-L: \*\*27.80\*\*

* Gigaword

+ ROUGE-1: \*\*36.30\*\*
+ ROUGE-2: 17.31
+ ROUGE-L: \*\*33.88\*\*

| NOT YET AVAILABLE | 2017
[Convolutional Sequence to Sequence](https://arxiv.org/pdf/1705.03122.pdf) |

* DUC-2004
* Gigaword

|

* DUC-2004

+ ROUGE-1: 33.44
+ ROUGE-2: \*\*10.84\*\*
+ ROUGE-L: 26.90

* Gigaword

+ ROUGE-1: 35.88
+ ROUGE-2: 27.48
+ ROUGE-L: 33.29

| [PyTorch](https://github.com/facebookresearch/fairseq-py) | 2017
#### 8. 의존 뼜 구문 분석 (Dependency Parsing)
연구 논문 | 데이터셋 | 지표 | 소스 코드 | 연도
------------ | ------------- | ------------ | ------------- | -------------
[Globally Normalized Transition-Based Neural Networks](https://arxiv.org/pdf/1603.06042.pdf) |

* 최종 CoNLL ’09 의존 뼜 구문 분석

|

* 94.08% UAS 정확도
* 92.15% LAS 정확도

|

* [SyntaxNet](https://github.com/tensorflow/models/tree/master/research/syntaxnet)

|

* 2017

### 컴퓨터 비전 (Computer Vision)
#### 1. 분류 (Classification)

| 연구 논문 | 데이터셋 | 지표 | 소스 코드 | 연도 |
| --- | --- | --- | --- | --- |
| [Dynamic Routing Between Capsules](https://arxiv.org/pdf/1710.09829.pdf) | * MNIST | * 테스트 오류: 0.25±0.005 | * [Official Implementation](https://github.com/Sarasra/models/tree/master/research/capsules) * [PyTorch](https://github.com/gram-ai/capsule-networks) * [Tensorflow](https://github.com/naturomics/CapsNet-Tensorflow) * [Keras](https://github.com/XifengGuo/CapsNet-Keras) * [Chainer](https://github.com/soskek/dynamic_routing_between_capsules) * [List of all implementations](https://github.com/loretoparisi/CapsNet) | 2017 |
| [High-Performance Neural Networks for Visual Object Classification](https://arxiv.org/pdf/1102.0183.pdf) | * NORB | * 테스트 오류: 2.53 ± 0.40 | * 찾을 수 없음 | 2011 |
| [Giant AmoebaNet with GPipe](https://arxiv.org/pdf/1811.06965.pdf) | * CIFAR-10 * CIFAR-100 * ImageNet-1k * ... | * 테스트 오류: 1.0% * 테스트 오류: 8.7% * Top-1 오류 15.7 * ... | * 찾을 수 없음 | 2018 |
| [ShakeDrop regularization](https://openreview.net/pdf?id=S1NHaMW0b) | * CIFAR-10 * CIFAR-100 | * 테스트 오류: 2.31% * 테스트 오류: 12.19% | * 찾을 수 없음 | 2017 |
| [Aggregated Residual Transformations for Deep Neural Networks](https://arxiv.org/pdf/1611.05431.pdf) | * CIFAR-10 | * 테스트 오류: 3.58% | * [PyTorch](https://github.com/facebookresearch/ResNeXt) | 2017 |
| [Random Erasing Data Augmentation](https://arxiv.org/abs/1708.04896) | * CIFAR-10 * CIFAR-100 * Fashion-MNIST | * 테스트 오류: 3.08% * 테스트 오류: 17.73% * 테스트 오류: 3.65% | [Pytorch](https://github.com/zhunzhong07/Random-Erasing) | 2017 |
| [EraseReLU: A Simple Way to Ease the Training of Deep Convolution Neural Networks](https://arxiv.org/abs/1709.07634) | * CIFAR-10 * CIFAR-100 | * 테스트 오류: 3.56% * 테스트 오류: 16.53% | [Pytorch](https://github.com/D-X-Y/EraseReLU) | 2017 |
| [Dynamic Routing Between Capsules](https://arxiv.org/pdf/1710.09829.pdf) | * MultiMNIST | * 테스트 오류: 5% | * [PyTorch](https://github.com/gram-ai/capsule-networks) * [Tensorflow](https://github.com/naturomics/CapsNet-Tensorflow) * [Keras](https://github.com/XifengGuo/CapsNet-Keras) * [Chainer](https://github.com/soskek/dynamic_routing_between_capsules) * [List of all implementations](https://github.com/loretoparisi/CapsNet) | 2017 |
| [Learning Transferable Architectures for Scalable Image Recognition](https://arxiv.org/pdf/1707.07012.pdf) | * ImageNet-1k | * Top-1 오류: 17.3 | * [Tensorflow](https://github.com/tensorflow/models/tree/master/research/slim/nets/nasnet) | 2017 |
| [Squeeze-and-Excitation Networks](https://arxiv.org/pdf/1709.01507.pdf) | * ImageNet-1k | * Top-1 오류: 18.68 | * [CAFFE](https://github.com/hujie-frank/SENet) | 2017 |
| [Aggregated Residual Transformations for Deep Neural Networks](https://arxiv.org/pdf/1611.05431.pdf) | * ImageNet-1k | * Top-1 오류: 20.4% | * [Torch](https://github.com/facebookresearch/ResNeXt) | 2016 |

#### 2. 인스턴스 분할 (Instance Segmentation)

| Research Paper | Datasets | Metric | Source Code | Year |
| --- | --- | --- | --- | --- |
| [Mask R-CNN](https://arxiv.org/pdf/1703.06870.pdf) | * COCO | * Average Precision: 37.1% | * [Detectron (Official Version)](https://github.com/facebookresearch/Detectron) * [MXNet](https://github.com/TuSimple/mx-maskrcnn) * [Keras](https://github.com/matterport/Mask_RCNN) * [TensorFlow](https://github.com/CharlesShang/FastMaskRCNN) | 2017 |

#### 3. Visual Question Answering

| Research Paper | Datasets | Metric | Source Code | Year |
| --- | --- | --- | --- | --- |
| [Tips and Tricks for Visual Question Answering: Learnings from the 2017 Challenge](https://arxiv.org/abs/1708.02711) | * VQA | * Overall score: 69 | * NOT FOUND | 2017 |

#### 4. Person Re-identification

| Research Paper | Datasets | Metric | Source Code | Year |
| --- | --- | --- | --- | --- |
| [Random Erasing Data Augmentation](https://arxiv.org/abs/1708.04896) | * [Market-1501](http://www.liangzheng.org/Project/project_reid.html) * [CUHK03-new-protocol](https://github.com/zhunzhong07/person-re-ranking) * [DukeMTMC-reID](https://github.com/layumi/DukeMTMC-reID_evaluation) | * Rank-1: 89.13 mAP: 83.93 * Rank-1: 84.02 mAP: 78.28 * labeled (Rank-1: 63.93 mAP: 65.05) detected (Rank-1: 64.43 mAP: 64.75) | [Pytorch](https://github.com/zhunzhong07/Random-Erasing) | 2017 |

### Speech
[Speech SOTA](https://github.com/syhw/wer\_are\_we)
#### 1. ASR

| Research Paper | Datasets | Metric | Source Code | Year |
| --- | --- | --- | --- | --- |
| [The Microsoft 2017 Conversational Speech Recognition System](https://arxiv.org/pdf/1708.06073.pdf) | * Switchboard Hub5'00 | * WER: 5.1 | * NOT FOUND | 2017 |
| [The CAPIO 2017 Conversational Speech Recognition System](https://arxiv.org/pdf/1801.00059.pdf) | * Switchboard Hub5'00 | * WER: 5.0 | * NOT FOUND | 2017 |

## Semi-supervised Learning
#### Computer Vision

| Research Paper | Datasets | Metric | Source Code | Year |
| --- | --- | --- | --- | --- |
| [DISTRIBUTIONAL SMOOTHINGWITH VIRTUAL ADVERSARIAL TRAINING](https://arxiv.org/pdf/1507.00677.pdf) | * SVHN * NORB | * Test error: 24.63 * Test error: 9.88 | [Theano](https://github.com/takerum/vat) | 2016 |
| [Virtual Adversarial Training: a Regularization Method for Supervised and Semi-supervised Learning](https://arxiv.org/pdf/1704.03976.pdf) | * MNIST | * Test error: 1.27 | * NOT FOUND | 2017 |
| [Few Shot Object Detection](https://arxiv.org/pdf/1706.08249.pdf) | * VOC2007 * VOC2012 | * mAP : 41.7 * mAP : 35.4 | * NOT FOUND | 2017 |
| [Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitro](https://arxiv.org/pdf/1701.07717.pdf) | * [Market-1501](http://www.liangzheng.org/Project/project_reid.html) * CUHK-03 * [DukeMTMC-reID](https://github.com/layumi/DukeMTMC-reID_evaluation) * [CUB-200-2011](http://www.vision.caltech.edu/visipedia/CUB-200-2011.html) | * Rank-1: 83.97 mAP: 66.07 * Rank-1: 84.6 mAP: 87.4 * Rank-1: 67.68 mAP: 47.13 * Test Accuracy: 84.4 | [Matconvnet](https://github.com/layumi/Person-reID_GAN) | 2017 |

## Unsupervised Learning
#### Computer Vision
##### 1. Generative Model

| Research Paper | Datasets | Metric | Source Code | Year |
| --- | --- | --- | --- | --- |
| [PROGRESSIVE GROWING OF GANS FOR IMPROVED QUALITY, STABILITY, AND VARIATION](http://research.nvidia.com/sites/default/files/publications/karras2017gan-paper-v2.pdf) | Unsupervised CIFAR 10 | Inception score: 8.80 | [Theano](https://github.com/tkarras/progressive_growing_of_gans) | 2017 |

### NLP
#### Machine Translation

|  |  |  |  |  |  |  |  |  |  |  |  |  |  |  |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 연구 논문 | 데이터셋 | 평가지표 | 소스 코드 | 연도 || [UNSUPERVISED MACHINE TRANSLATION USING MONOLINGUAL CORPORA ONLY](https://arxiv.org/pdf/1711.00043.pdf) | * Multi30k-Task1(en-fr fr-en de-en en-de) | * BLEU:(32.76 32.07 26.26 22.74) | * 찾을 수 없음 | 2017 |
| [Unsupervised Neural Machine Translation with Weight Sharing](https://arxiv.org/pdf/1804.09057.pdf) | * WMT14(en-fr fr-en) * WMT16 (de-en en-de) | * BLEU:(16.97 15.58) * BLEU:(14.62 10.86) | * 찾을 수 없음 | 2018 |

## 전이 학습 (Transfer Learning)

|  |  |  |  |  |  |  |  |  |  |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 연구 논문 | 데이터셋 | 평가지표 | 소스 코드 | 연도 || [One Model To Learn Them All](https://arxiv.org/pdf/1706.05137.pdf) | * WMT EN → DE * WMT EN → FR (BLEU) * ImageNet (top-5 accuracy) | * BLEU: 21.2 * BLEU:30.5 * 86% | * [Tensorflow](https://github.com/tensorflow/tensor2tensor) | 2017 |

## 강화 학습 (Reinforcement Learning)

|  |  |  |  |  |  |  |  |  |  |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 연구 논문 | 데이터셋 | 평가지표 | 소스 코드 | 연도 || [Mastering the game of Go without human knowledge](http://www.gwern.net/docs/rl/2017-silver.pdf) | 바둑 | Elo 등급: 5185 | * [C++](https://github.com/gcp/leela-zero) | 2017 |

이메일: yxt.stoaml@gmail.com

state-of-the-art-result-for-machine-learning-problems

파일 탐색기

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

state-of-the-art-result-for-machine-learning-problems

파일 탐색기

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README