I'd first of all like to express my gratitude for the Kaggle team to organize this challenge and my colleagues from my office for supporting me through the journey.
\nApart from the coding & conceptual knowledge that I gained through this I also learnt the most important values of patience, perseverence & hardwork.
\nLate into the competition I also joined in the efficiency track, so that was an additional learning i got about making the inference superfast and actually developing strong & meaningful textual features.
this is my code for the efficiency track: https://www.kaggle.com/bhavesjain/commonlit-2023-train
\nand for the accuracy track i modified the publicly available solution with my fixes: https://www.kaggle.com/code/bhavesjain/commonlit-ess-lgbm-autocorrect-deberta-v3-tuned/
\nOverview:
\nMy prime objective in the journey was to experiment with logical approaches than just bruteforce techniques of concatenating everything & feeding it to a powerful & large models. I experimented with textual features like lexical diversity, readability indices, syllables& word metrics. I wanted to explore more into the area of metadiscourse & publicly available datasets for grading the quality but i had time & computation quota limitations. I also experimented with BERT, different versions of DeBERTa, sentence transformers but that's about it.
\nSubmission Details:
\nI primarily stuck to XGBoost with Optuna. And rest was just hit & trial with various models.
\nThe USP of my solution is the wide array of features that i used inspired from https://textinspector.com/help/statistics-readability/
\nI hope you all too check it out
\nAdditionally i also learnt that i need to focus more on the robustness of my training techniques & optimization to avoid such loss in shakeups.
Even though I didn't get a medal, i'm glad that i learnt a lot of things worth applying in my own domain & possibly in upcoming competitions.
", "comments": [ { "id": 2479537, "author_name": "Archana Jain", "author_url": "", "post_date": "2023-10-12T17:14:44.373000", "content": "That's a nice angle to tackle the competition without even any hardcore deep learning
", "votes": 1, "replies": [] } ], "comment_count": 1, "messages": [ { "id": 2478690, "postDate": "2023-10-12T06:42:57.690Z", "content": "I'd first of all like to express my gratitude for the Kaggle team to organize this challenge and my colleagues from my office for supporting me through the journey.
\nApart from the coding & conceptual knowledge that I gained through this I also learnt the most important values of patience, perseverence & hardwork.
\nLate into the competition I also joined in the efficiency track, so that was an additional learning i got about making the inference superfast and actually developing strong & meaningful textual features.
this is my code for the efficiency track: https://www.kaggle.com/bhavesjain/commonlit-2023-train
\nand for the accuracy track i modified the publicly available solution with my fixes: https://www.kaggle.com/code/bhavesjain/commonlit-ess-lgbm-autocorrect-deberta-v3-tuned/
\nOverview:
\nMy prime objective in the journey was to experiment with logical approaches than just bruteforce techniques of concatenating everything & feeding it to a powerful & large models. I experimented with textual features like lexical diversity, readability indices, syllables& word metrics. I wanted to explore more into the area of metadiscourse & publicly available datasets for grading the quality but i had time & computation quota limitations. I also experimented with BERT, different versions of DeBERTa, sentence transformers but that's about it.
\nSubmission Details:
\nI primarily stuck to XGBoost with Optuna. And rest was just hit & trial with various models.
\nThe USP of my solution is the wide array of features that i used inspired from https://textinspector.com/help/statistics-readability/
\nI hope you all too check it out
\nAdditionally i also learnt that i need to focus more on the robustness of my training techniques & optimization to avoid such loss in shakeups.
Even though I didn't get a medal, i'm glad that i learnt a lot of things worth applying in my own domain & possibly in upcoming competitions.
", "rawMarkdown": "I'd first of all like to express my gratitude for the Kaggle team to organize this challenge and my colleagues from my office for supporting me through the journey.\nApart from the coding & conceptual knowledge that I gained through this I also learnt the most important values of patience, perseverence & hardwork.\nLate into the competition I also joined in the efficiency track, so that was an additional learning i got about making the inference superfast and actually developing strong & meaningful textual features.\n\nthis is my code for the efficiency track: https://www.kaggle.com/bhavesjain/commonlit-2023-train\n\nand for the accuracy track i modified the publicly available solution with my fixes: https://www.kaggle.com/code/bhavesjain/commonlit-ess-lgbm-autocorrect-deberta-v3-tuned/\nOverview: \nMy prime objective in the journey was to experiment with logical approaches than just bruteforce techniques of concatenating everything & feeding it to a powerful & large models. I experimented with textual features like lexical diversity, readability indices, syllables& word metrics. I wanted to explore more into the area of metadiscourse & publicly available datasets for grading the quality but i had time & computation quota limitations. I also experimented with BERT, different versions of DeBERTa, sentence transformers but that's about it.\nSubmission Details:\nI primarily stuck to XGBoost with Optuna. And rest was just hit & trial with various models.\nThe USP of my solution is the wide array of features that i used inspired from https://textinspector.com/help/statistics-readability/\nI hope you all too check it out\nAdditionally i also learnt that i need to focus more on the robustness of my training techniques & optimization to avoid such loss in shakeups.\n\nEven though I didn't get a medal, i'm glad that i learnt a lot of things worth applying in my own domain & possibly in upcoming competitions.", "votes": 9 }, { "id": 2479537, "postDate": "2023-10-12T17:14:44.373Z", "content": "That's a nice angle to tackle the competition without even any hardcore deep learning
", "rawMarkdown": "That's a nice angle to tackle the competition without even any hardcore deep learning ", "votes": 1 } ], "complete": true, "source": "live", "fetched_at": "2026-10-19T23:02:05Z", "competition_id": 53482, "competition_slug": "commonlit-evaluate-student-summaries", "team_id": 10676814, "solution_url": "/competitions/commonlit-evaluate-student-summaries/writeups/bhavesh-jain-learnings-from-this-competition" }