0.25
0.5
0.75
1.25
1.5
1.75
2
Presentation
Natural Language Processing, Language Modelling and Machine Translation00:00
Natural Language Processing11:14:40
Language models60:23:09
History: cryptography93:30:16
Language models - 1103:47:10
Language models - 2124:03:59
Language models - 3140:38:05
Evaluating a Language Model169:01:17
Language Modelling Data - 1183:11:26
Language Modelling Data - 2196:50:23
Language Modelling Overview230:41:47
N-Gram Models: The Markov Chain Assumption247:37:33
N-Gram Models: Estimating Probabilities260:32:33
N-Gram Models: Back-O 269:13:27
N-Gram Models: Interpolated Back-O 287:52:19
Provisional Summary313:21:28
Outline - 1365:49:32
Neural Language Models - 1367:14:07
Neural Language Models - 2372:07:30
Neural Language Models: Sampling - 1395:00:05
Neural Language Models: Sampling - 2399:49:09
Neural Language Models: Training - 1404:47:00
Neural Language Models: Training - 2408:47:40
Neural Language Models: Training - 3409:17:01
Comparison with Count Based N-Gram LMs432:35:56
Recurrent Neural Network Language Models - 1455:14:18
Recurrent Neural Network Language Models - 2469:08:54
Recurrent Neural Network Language Models - 3477:17:37
Recurrent Neural Network Language Models - 4482:55:26
Recurrent Neural Network Language Models - 5489:48:01
Recurrent Neural Network Language Models - 6514:34:00
Recurrent Neural Network Language Models - 7531:00:24
Comparison with N-Gram LMs531:32:25
Language Modelling: Review565:39:26
Gated Units: LSTMs and GRUs583:31:10
Deep RNN LMs - 1593:30:10
Deep RNN LMs - 2595:21:01
Deep RNN LMs - 3596:03:56
Deep RNN LMs - 4607:57:07
Deep RNN LM - 1612:13:50
Deep RNN LM - 2614:35:39
Scaling: Large Vocabularies - 1627:08:26
Scaling: Large Vocabularies - 2670:06:33
Scaling: Large Vocabularies - 3681:06:59
Scaling: Large Vocabularies - 4700:33:44
Scaling: Large Vocabularies - 5708:48:22
Scaling: Large Vocabularies - 6727:20:56
Scaling: Large Vocabularies - 7750:42:13
Sub-Word Level Language Modelling772:49:39
Regularisation: Dropout - 1840:30:55
Regularisation: Dropout - 2846:25:34
Regularisation: Bayesian Dropout (Gal)864:50:55
Evaluation: hyperparamters are a confounding factor876:09:29
Summary1046:39:17
Intro to MT1054:30:49
Parallel Corpora1071:48:20
MT History: Statistical MT at IBM - 11091:32:32
MT History: Statistical MT at IBM - 21116:06:41
Models of translation - 11135:42:25
Models of translation - 21157:32:40
IBM Model 1: The first translation attention model!1161:56:31
Encoder-Decoders131180:46:53
Recurrent Encoder-Decoders for MT14 - 11194:38:02
Recurrent Encoder-Decoders for MT14 - 21207:05:27
Recurrent Encoder-Decoders for MT14 - 31217:24:12
Attention Models for MT15 - 11218:23:33
Attention Models for MT15 - 21230:55:37
Attention Models for MT15 - 31234:36:01
Attention Models for MT15 - 41236:40:11
Attention Models for MT15 - 51246:42:33
Attention Models for MT15 - 61248:25:01
Returning to the Noisy Channel - 11262:30:16
Returning to the Noisy Channel - 21291:29:52
Decoding1310:58:46
Decoding: Direct vs. Noisy Channel - 11316:32:06
Decoding: Direct vs. Noisy Channel - 21323:25:49
Decoding: Noisy Channel Model1332:05:33
Segment to Segment Neural Transduction1345:11:36
Noisy Channel Decoding1373:38:34
Relative Performance161387:13:01
The End1401:51:59