{"id":19544166,"url":"https://github.com/sunset1995/talk2ai_contest","last_synced_at":"2025-04-26T17:32:57.244Z","repository":{"id":85799119,"uuid":"100114714","full_name":"sunset1995/Talk2AI_Contest","owner":"sunset1995","description":"科技大擂台 與AI對話","archived":false,"fork":false,"pushed_at":"2017-11-01T06:51:53.000Z","size":13898,"stargazers_count":6,"open_issues_count":0,"forks_count":0,"subscribers_count":5,"default_branch":"master","last_synced_at":"2025-04-04T17:01:50.764Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/sunset1995.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2017-08-12T13:41:10.000Z","updated_at":"2022-06-20T14:18:11.000Z","dependencies_parsed_at":null,"dependency_job_id":"bb8249eb-f3ef-48a9-8911-f83e6df4c2ec","html_url":"https://github.com/sunset1995/Talk2AI_Contest","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sunset1995%2FTalk2AI_Contest","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sunset1995%2FTalk2AI_Contest/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sunset1995%2FTalk2AI_Contest/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/sunset1995%2FTalk2AI_Contest/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/sunset1995","download_url":"https://codeload.github.com/sunset1995/Talk2AI_Contest/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":251026576,"owners_count":21525001,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2024-11-11T03:24:55.155Z","updated_at":"2025-04-26T17:32:57.238Z","avatar_url":"https://github.com/sunset1995.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# 科技部科技大擂台 與AI對話 熱身賽\n\n隊名：現充與現充與魯蛇\n隊員：[ChiWeiHsiao](https://github.com/ChiWeiHsiao), [sunset](https://github.com/sunset1995), [SouthRa](https://github.com/ko19951231)\n\n我們嘗試過的方法:\n- `naive`\n    - count 下句的詞在上句出現幾次\n    - 將一句話 word2vec 後的 vector 取(加權)平均，算上下句的 cosine similarity\n- `dual-lstm`\n    - 兩個不同的 rnn encode 上下句後，`sigmoid(u, Mv)` 來 train 其認識上下句\n    - 用 `sigmoid(u, Mv)` 來代表為上下句的機率\n    - 用 `u`, `v` 的 cosine similarity\n- `rnn-encoder`\n    - 同 `dual-lstm` 但上下句的 rnn 為 shared weights\n- `smn`\n    - [Sequential Matching Network: A New Architecture for Multi-turn\nResponse Selection in Retrieval-Based Chatbots](https://arxiv.org/pdf/1612.01627.pdf)\n\n最後版本：選 variance 高的多個 model 的答案做 voting\n- dual-lstm (dual-lstm 的 rnn-encoder 差異性較小，但 dual-lstm 表現較好故選之)\n- 多個 train deep model 時 fine-tuned 過的 word2vec 跑 naive 方法，總票數比 dual-lstm 少\n- smn 沒時間 tune 跟 train\n\n\n熱身賽決賽預賽結果：  \n\n| 隊伍 | 語音分數(現場運氣) | 文本分數 | 總分 |\n| --- | :--- | :--- | :--- |\n| 現充與現充與魯蛇 | 0.6756 | 0.5810 | 0.6094 |\n| other team | 0.7567 | 0.5185 | 0.5899 |\n| other team | 0.5945 | 0.4955 | 0.5252 |\n| other team | 0.5945 | 0.4890 | 0.5206 |\n| other team | 0.6216 | 0.4725 | 0.5172 |\n| other team | 0.5945 | 0.4695 | 0.5070 |\n| other team | 0.4864 | 0.4780 | 0.4805 |\n| other team | 0.4864 | 0.4565 | 0.4654 |\n| other team | 0.4864 | 0.4500 | 0.4609 |\n| other team | 0.5135 | 0.4320 | 0.4564 |\n\n\n熱身賽決賽四強結果：  \n\n| 隊伍 | 語音分數(現場運氣) | 文本分數 | 總分 |\n| --- | :--- | :--- | :--- |\n| 現充與現充與魯蛇 | 0.6101 | 0.5875 | 0.5943 |\n| other team | 0.4915 | 0.5190 | 0.5107 |\n| other team | 0.4915 | 0.5105 | 0.5048 |\n| other team | 0.4237 | 0.5135 | 0.4865 |\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsunset1995%2Ftalk2ai_contest","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsunset1995%2Ftalk2ai_contest","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsunset1995%2Ftalk2ai_contest/lists"}