Following up on the last topic about Textual similarity approach to solve the problem, one another potential approach I was thinking that may be helpful for this problem is Textual Entailment based methods.
\nFrom the Wikipedia page about textual entailment,
\nIn the Textual Entailment framework, the entailing and entailed texts are termed text (t) and hypothesis (h), respectively. \nTextual entailment is not the same as pure logical entailment – it has a more relaxed definition: \"t entails h\" (t ⇒ h) if, typically, a human reading t would infer that h is most likely true. (Alternatively: t ⇒ h if and only if, typically, a human reading t would be justified in inferring the proposition expressed by h from the proposition expressed by t)\nThe textual entailment objective for our case would be whether markdown cell entails the code cell or vice versa. Once we build the models, we can rank-order the pairs based on the entailment score and then select the top markdown cell as the next / previous cell.
Some challenges were conducted for Textual Entailment in the past and can be seen in
\nhttps://aclweb.org/aclwiki/Recognizing_Textual_Entailment
Some of the top approaches for recognizing textual entailment are present in
\nhttps://paperswithcode.com/dataset/rte
Further references:
\n\nPlease let me know your thoughts. Thank you.
", "messages": [ { "id": 1792589, "postDate": "2022-05-17T05:38:19.663Z", "content": "Following up on the last topic about Textual similarity approach to solve the problem, one another potential approach I was thinking that may be helpful for this problem is Textual Entailment based methods.
\nFrom the Wikipedia page about textual entailment,
\nIn the Textual Entailment framework, the entailing and entailed texts are termed text (t) and hypothesis (h), respectively. \nTextual entailment is not the same as pure logical entailment – it has a more relaxed definition: \"t entails h\" (t ⇒ h) if, typically, a human reading t would infer that h is most likely true. (Alternatively: t ⇒ h if and only if, typically, a human reading t would be justified in inferring the proposition expressed by h from the proposition expressed by t)\nThe textual entailment objective for our case would be whether markdown cell entails the code cell or vice versa. Once we build the models, we can rank-order the pairs based on the entailment score and then select the top markdown cell as the next / previous cell.
Some challenges were conducted for Textual Entailment in the past and can be seen in
\nhttps://aclweb.org/aclwiki/Recognizing_Textual_Entailment
Some of the top approaches for recognizing textual entailment are present in
\nhttps://paperswithcode.com/dataset/rte
Further references:
\n\nPlease let me know your thoughts. Thank you.
", "rawMarkdown": "Following up on the last topic about [Textual similarity approach](https://www.kaggle.com/competitions/AI4Code/discussion/324509) to solve the problem, one another potential approach I was thinking that may be helpful for this problem is [Textual Entailment based methods](https://en.wikipedia.org/wiki/Textual_entailment).\n\nFrom the Wikipedia page about textual entailment,\n```\nIn the Textual Entailment framework, the entailing and entailed texts are termed text (t) and hypothesis (h), respectively. \nTextual entailment is not the same as pure logical entailment – it has a more relaxed definition: \"t entails h\" (t ⇒ h) if, typically, a human reading t would infer that h is most likely true. (Alternatively: t ⇒ h if and only if, typically, a human reading t would be justified in inferring the proposition expressed by h from the proposition expressed by t)\n```\n\nThe textual entailment objective for our case would be whether `markdown cell` entails the `code cell` or vice versa. Once we build the models, we can rank-order the pairs based on the entailment score and then select the top markdown cell as the next / previous cell.\n\nSome challenges were conducted for Textual Entailment in the past and can be seen in\nhttps://aclweb.org/aclwiki/Recognizing_Textual_Entailment\n\nSome of the top approaches for recognizing textual entailment are present in \nhttps://paperswithcode.com/dataset/rte\n\nFurther references:\n* [Textual Entailment References](https://www.cs.upc.edu/~ageno/anlp/textualEntailment.pdf)\n* [SNLI corpus](https://nlp.stanford.edu/projects/snli/)\n\nPlease let me know your thoughts. Thank you.", "votes": 20 } ], "comments": [], "raw_markdown_by_id": { "1792589": "Following up on the last topic about [Textual similarity approach](https://www.kaggle.com/competitions/AI4Code/discussion/324509) to solve the problem, one another potential approach I was thinking that may be helpful for this problem is [Textual Entailment based methods](https://en.wikipedia.org/wiki/Textual_entailment).\n\nFrom the Wikipedia page about textual entailment,\n```\nIn the Textual Entailment framework, the entailing and entailed texts are termed text (t) and hypothesis (h), respectively. \nTextual entailment is not the same as pure logical entailment – it has a more relaxed definition: \"t entails h\" (t ⇒ h) if, typically, a human reading t would infer that h is most likely true. (Alternatively: t ⇒ h if and only if, typically, a human reading t would be justified in inferring the proposition expressed by h from the proposition expressed by t)\n```\n\nThe textual entailment objective for our case would be whether `markdown cell` entails the `code cell` or vice versa. Once we build the models, we can rank-order the pairs based on the entailment score and then select the top markdown cell as the next / previous cell.\n\nSome challenges were conducted for Textual Entailment in the past and can be seen in\nhttps://aclweb.org/aclwiki/Recognizing_Textual_Entailment\n\nSome of the top approaches for recognizing textual entailment are present in \nhttps://paperswithcode.com/dataset/rte\n\nFurther references:\n* [Textual Entailment References](https://www.cs.upc.edu/~ageno/anlp/textualEntailment.pdf)\n* [SNLI corpus](https://nlp.stanford.edu/projects/snli/)\n\nPlease let me know your thoughts. Thank you." } }