“Tianyi Luo 2016-04-25”版本间的差异
来自cslt Wiki
| （相同用户的4个中间修订版本未显示） | |||
| 第16行： | 第16行： | ||
--------------------2016-04-22 | --------------------2016-04-22 | ||
* Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN. | * Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN. | ||
| + | * Help conduct the presentation of text group(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary. | ||
--------------------2016-04-23 | --------------------2016-04-23 | ||
* Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%. | * Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%. | ||
| 第21行： | 第22行： | ||
* Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72.09%. | * Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72.09%. | ||
* Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 72.09% to 72.86%. | * Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 72.09% to 72.86%. | ||
| − | * Use entity match rules 5(machine number + English chacracters) to improve the accuracy from 72.86% to | + | * Use entity match rules 5(machine number + English chacracters) to improve the accuracy from 72.86% to 76.48%. |
| + | * Use simlar pair match rules 1(“快”和“速度”、“复写”和“拷贝”等等) to improve the accuracy from 76.48% to 82.39%. | ||
=== Plan to do next week === | === Plan to do next week === | ||
* To implement tensorflow version of RNN/LSTM Max margin vector training. | * To implement tensorflow version of RNN/LSTM Max margin vector training. | ||
2016年4月25日 (一) 00:36的最后版本
Plan to do this week
- To implement tensorflow version of RNN/LSTM Max margin vector training.
Work done in this week
2016-04-18
- Optimize theano version of Generationg the similar questions' vectors based on RNN.
- Finish implementing theano version of LSTM Max margin vector training.
2016-04-19
- Optimize theano version of Generationg the similar questions' vectors based on RNN.
2016-04-20
- Finish submiting the camera version paper of IJCAI 2016.
- Update the version of Technical Report about Chinese Song Iambics generation.
2016-04-21
- Finish helping Teacher Wang to prepare for text group's presentation(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.
- Submit our IJCAI paper to arxiv. (Solve a big problem about submitting the paper including Chinese chacracters. Solution)
- Optimize theano version of Generationg the similar questions' vectors based on RNN.
2016-04-22
- Speed up process of the test performance about theano version of Generationg the similar questions' vectors based on RNN.
- Help conduct the presentation of text group(Tang poetry and Songci generation and Intelligent QA system) for Tsinghua University's 105 anniversary.
2016-04-23
- Use entity match rule 1(dpk, ix, lpk sv + machine number) to improve the accuracy from 38% to 58%.
2016-04-24
- Use entity match rules 2+3(dps, s, e, fp, tps, q, 200 + machine number and machine number only) to improve the accuracy from 58% to 72.09%.
- Use entity match rules 4(dpk, ix, lpk sv, dps, s, e, fp, tps, q + machine number + English chacracters) to improve the accuracy from 72.09% to 72.86%.
- Use entity match rules 5(machine number + English chacracters) to improve the accuracy from 72.86% to 76.48%.
- Use simlar pair match rules 1(“快”和“速度”、“复写”和“拷贝”等等) to improve the accuracy from 76.48% to 82.39%.
Plan to do next week
- To implement tensorflow version of RNN/LSTM Max margin vector training.
- To implement attention chatting model with xiaobing corpus.
Interested papers
- Cascading Bandits: Learning to Rank in the Cascade Model(ICML 2015) [pdf]
