| MO810/MC959 | 2018 First Semester | Unicamp, Brazil | ||||||||||||||||||
| 5.00% | 5.00% | 5.00% | ||||||||||||||||||
| 04/03/18 | 04/03/18 | Topic, Dataset | 05/03/18 | 05/03/18 | Model, Baseline | 06/14/18 | 06/14/18 | 06/14/18 | 06/14/18 | 06/14/18 | ||||||||||
| N | Group | Topic | Dataset | Dataset (with URL if available) | Delivery | Penalty | Model | Performance Measure | Baseline | Delivery | Penalty | Report | Delivery | Penalty | Video | Delivery | Penalty | Relevance | Difficulty | Compliance |
| 1 | Ishikawa | Porn in cartoons | private; hundreds of videos | 2018-04-01 22:03:07 | 0.00% | GoogLeNet+softmax;SVM | F2 and normalized accuracy | 80%,90%,91%; 87%,93%,94% | 2018-05-03 20:37:55 | 0.00% | Cartoon Porn Detection Through Deep Learning and Motion Information | 2018-06-14 12:52:22 | 0.00% | Cartoon Porn Detection Through Deep Learning and Motion Information | 2018-06-14 12:52:22 | 0.00% | 10 | 10 | 10 | |
| 2 | Gibaut | Music recommendation | grouplens hetrec2011 Last.FM | https://grouplens.org/datasets/hetrec-2011/, Last.FM | 2018-04-02 21:26:00 | 0.00% | autoencoder: 18745;500;18745 | MSE | 1.0 to 1.5 | 2018-05-03 21:35:33 | 0.00% | A Music Recommendation System with individual and social approach | 2018-06-14 18:37:28 | 0.00% | (diff title) | 2018-06-14 18:37:28 | 0.00% | 10 | 10 | 10 |
| 3 | Padovani-Santos | GAN (?) | MNIST, CIFAR-10 | dropped | dropped | |||||||||||||||
| 4 | Froes | Circuit quality assessment | Customer call data | 2018-04-07 03:35:21 | -62.99% | dropped | ||||||||||||||
| 5 | Camargo | Noise in Audio | Research Google Audioset | https://research.google.com/audioset/dataset/index.html | 2018-04-02 22:33:58 | 0.00% | 15-layer convolutional net; description in email | 2018-05-03 16:10:32 | 0.00% | Deep Convolutional Neural Network for Audio Classification | 2018-06-14 21:11:09 | 0.00% | (no title) | 2018-06-14 21:11:09 | 0.00% | 10 | 10 | 10 | ||
| 6 | Hasegawa | Stock Market | B3 ftp site, Perlin | ftp://ftp.bmf.com.br/MarketData/BMF/ | 2018-04-03 12:48:41 | 0.00% | PG, 1 layer; V, 3 layers; env.modif.from Trading Gym (https://github.com/Prediction-Machines/Trading-Gym) | # of positive operations; total reward | 240; -0.2 points | 2018-05-03 22:43:59 | 0.00% | (no title) | 2018-06-14 23:06:12 | 0.00% | Automated Trading with Reinforcement Learning | 2018-06-14 23:06:12 | 0.00% | 10 | 10 | 10 |
| 7 | Aguiar-Capone-Sangalli | Sentiment analysis | Book reviews, Julian McAuley | http://snap.stanford.edu/data/amazon/productGraph/categoryFiles/review s_Books_5.json.gz | 2018-04-03 13:56:54 | 0.00% | word emb 100;[ch=128,k=3,mp=3;ch=128mp=3,k=4;ch=128,k=5,mp=3];ch=128,k=5,mp=5;ch=128,k=5,mp=5;ch=128,k=5,mp=20; (+? LSTM) | accuracy on fine-grained (5-item) classes | 53% just CNN; 53% paper | 2018-05-03 21:50:11 | 0.00% | Sentiment analysis of Amazon’s reviews using deep learning | 2018-06-14 20:16:13 | 0.00% | Sentiment analysis of Amazon’s reviews using deep learning | 2018-06-14 20:16:13 | 0.00% | 10 | 10 | 10 |
| 8 | Ricci-Claus | Handwritten digits | MNIST | http://yann.lecun.com/exdb/mnist/ | 2018-04-03 17:47:21 | 0.00% | goal: smallest networks achieving 95% and 99% | accuracy | ch=32,k=5;maxp=2;ch=64,k=3;maxp=2;d=1024,dr=0.25 | 2018-05-03 16:52:01 | 0.00% | Handwritten Digits Recognition | 2018-06-14 22:02:04 | 0.00% | Handwritten Digits Recognition | 2018-06-14 22:02:04 | 0.00% | 10 | 0 | 10 |
| 9 | Campos-Costa | Stock Market | Kaggle: daily US stock info | https://www.kaggle.com/borismarjanovic/price-volume-data-for-all-us-stocks-etfs | 2018-04-03 18:05:26 | 0.00% | profitability (%/day) | 0.05 | 2018-05-03 20:14:18 | 0.00% | Stock Market Prediction and Trader: a machine learning approach | 2018-06-14 23:51:33 | 0.00% | Stock Market Prediction and Trader: a machine learning approach | 2018-06-14 23:51:33 | 0.00% | 10 | 10 | 10 | |
| 10 | Lopes-Moreira | Image classification | Dog Breeds | http://vision.stanford.edu/aditya86/ImageNetDogs/ | 2018-04-03 19:12:03 | 0.00% | Xception+InceptionV3;d=128,L2;d=128,L2;softmax | accuracy, cross-entropy loss | 88.8%, 0.40 | 2018-05-03 16:55:04 | 0.00% | Dog Breed Classification using Transfer Learning | 2018-06-15 00:03:16 | -0.05% | Dog Breed Classification using Transfer Learning | 2018-06-15 00:03:16 | -0.05% | 10 | 10 | 10 |
| 11 | Tosi-Borges-Malta | Word prediction | Kaggle MLSD - HW3 | https://www.kaggle.com/c/mlsd-hw3 | 2018-04-03 22:21:47 | 0.00% | word embedding (300), 2-layer LSTM, 450 units/layer; dropout | accuracy | 45% DL (53% other) | 2018-05-03 21:45:40 | 0.00% | A deep learning approach to word prediction | 2018-06-14 21:59:24 | 0.00% | A deep learning approach to word prediction | 2018-06-14 21:59:24 | 0.00% | 10 | 10 | 10 |
| 12 | Bahia (Moreira-Franco) | Blackjack | Server/client environment | https://github.com/joaopedroffranco/blackjack, https://github.com/KIQ83/blackjack-client | 2018-04-03 19:30:43 | 0.00% | (1) d=50;softmax (2) d=50;d=30;softmax | win rate | 14%, 47%, 37%, 53% | 2018-05-03 23:32:22 | 0.00% | (no title) | 2018-06-14 22:44:52 | 0.00% | The Blackjack | 2018-06-14 22:44:52 | 0.00% | 10 | 10 | 10 |
| 13 | Robles | Evolutionary Pong | OpenAI Gym Pong-v0 | https://gym.openai.com/envs/#atari | 2018-04-03 19:46:27 | 0.00% | ch=32,k=8,str=4;ch=64,k=4,str=2;ch=64,k=3;d=512;d=18 | total reward | 18.4 | 2018-05-03 22:53:20 | 0.00% | Deep Reinforcement Learning to play Pong | 2018-06-14 22:58:56 | 0.00% | Deep Reinforcement Learning to play Pong | 2018-06-14 22:58:56 | 0.00% | 10 | 10 | 10 |
| 14 | Nagme | Chest X-rays | Kaggle NIH Chest X-rays | 2018-04-03 20:08:57 | 0.00% | dropped | ||||||||||||||
| 15 | Rodrigues | Hypothalamus | private; 200 MR images | private | 2018-04-03 20:30:22 | 0.00% | U-Net (Ronnenberg et al, 2015) | Dice index (F1 score, Sorensen-Dice coefficient) | 0.75 | 2018-05-03 23:58:07 | 0.00% | Hyphotalamus automatic Segmentation Using Convolutional Neural Networks | 2018-06-14 23:59:03 | 0.00% | Hyphotalamus automatic Segmentation Using Convolutional Neural Networks | 2018-06-14 23:59:03 | 0.00% | 10 | 10 | 10 |
| 16 | PacBot (Idalgo-Pavan) | Playing Pac Man | OpenAI Gym MsPacman-v0 | environment OpenAI Gym MsPacman-ram-v0 | 2018-04-03 23:39:25 | 0.00% | d=128;d=128;d=512;d=512;d=128; | OpenAI Gym reward | complete level | 2018-05-03 21:20:47 | 0.00% | PacBot - a Reinforcement Learning Report | 2018-06-15 23:41:02 | -19.74% | (no title) | 2018-06-15 23:41:02 | -19.74% | 10 | 10 | 10 |
| 17 | Silva-Prado | Soccer positions | (146,107)-sized images | MNIST, CIFAR-10, and private HullConvex; all grayscale (146,107) | 2018-04-03 21:37:23 | 0.00% | ch=3,k=3,str=3;ch=3,k=3,str=3;maxp=2,str=2;d=128;softmax | accuracy | 28.41% | 2018-05-06 17:35:45 | -54.66% | Image Classification using Transfer Learning | 2018-06-14 19:42:45 | 0.00% | (diff title) | 2018-06-14 19:42:45 | 0.00% | 10 | 10 | 10 |
| 18 | Chevrier-Bruno | Plants by leaves | Leafsnap | http://leafsnap.com/dataset/ | 2018-04-03 21:40:11 | 0.00% | ch=40,k=3;maxp=4,str=2;ch=40,k=2;maxp=4,str=2;ch=80,k=2;maxp=4,str=2;ch=160,k=2;maxp=2;d=500;softmax | accuracy | 80.00% | 2018-05-04 00:01:27 | -0.02% | AI learns to recognize plants by leaves | 2018-06-14 22:29:25 | 0.00% | AI learns to recognize plants by leaves | 2018-06-14 22:29:25 | 0.00% | 10 | 10 | 10 |
| 19 | Divino | Simple voice commands | TF Speech Recogn. Challenge | 2018-04-12 21:07:05 | -100.00% | (no delivery) | ||||||||||||||