https://github.com/csinva/csinva
readme
https://github.com/csinva/csinva
Last synced: 5 months ago
JSON representation
readme
- Host: GitHub
- URL: https://github.com/csinva/csinva
- Owner: csinva
- Created: 2020-10-23T05:21:22.000Z (almost 6 years ago)
- Default Branch: main
- Last Pushed: 2025-06-02T20:04:23.000Z (about 1 year ago)
- Last Synced: 2025-06-03T09:49:47.319Z (about 1 year ago)
- Size: 93.8 KB
- Stars: 1
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
Hi there 👋 I'm Chandan, a Senior Researcher at Microsoft Research working on interpretable machine learning.
Homepage / Twitter / Google Scholar / LinkedIn
## 🌳 Interpretable models / dataset explanations
**** Interpretable and accurate predictive modeling, sklearn-compatible ([JOSS 2021](https://joss.theoj.org/papers/10.21105/joss.03192)). Contains FIGS ([PNAS 2022](https://arxiv.org/abs/2201.11931)) and HSTree ([ICML 2022](https://arxiv.org/abs/2202.00858))
**** Interpretability for text. Contains Aug-imodels ([Nature Communications 2023](https://arxiv.org/abs/2209.11799)) , Tree-Prompt ([EMNLP 2023](https://arxiv.org/abs/2310.14034)) , iPrompt ([ICLR workshop 2023](https://arxiv.org/abs/2210.01848)) , SASC ([NeurIPS workshop 2023](https://arxiv.org/abs/2305.09863)) , and QA-Embs ([NeurIPS 2024](https://arxiv.org/abs/2405.16714)) 
**[adaptive-wavelets](https://github.com/Yu-Group/adaptive-wavelet-distillation) ** Adaptive, interpretable wavelets ([NeurIPS 2021](https://arxiv.org/abs/2107.09145))
## 🤖 General-purpose AI packages and cheatsheets
**** Notes and resources on AI
**** Utilities for trustworthy data-science ([JOSS 2021](https://joss.theoj.org/papers/10.21105/joss.03895))
## 🧠Interpreting neural networks
**[deep-explanation-penalization](https://github.com/laura-rieger/deep-explanation-penalization) ** Penalizing neural-network explanations ([ICML 2020](https://arxiv.org/abs/1909.13584))
**[hierarchical-dnn-interpretations](https://github.com/csinva/hierarchical-dnn-interpretations) ** Hierarchical interpretations for neural network predictions ([ICLR 2019](https://arxiv.org/abs/1806.05337))
**[transformation-importance](https://github.com/csinva/transformation-importance) ** Feature importance for transformations ([ICLR Workshop 2020](https://arxiv.org/abs/2003.01926))
## 📊 Data-science problems
**[automated-brain-explanations](https://github.com/microsoft/automated-brain-explanations) ** Building natural-language explanations for the brain. Contains GCT ([arxiv 2024](https://arxiv.org/abs/2410.00812))
**[clinical-rule-development](https://github.com/csinva/clinical-rule-development) ** Building and vetting clinical decision rules, including vetting an intraabdominal rule ([PLOS DH, 2022](https://journals.plos.org/digitalhealth/article?id=10.1371/journal.pdig.0000076)), analyzing patient perspectives for approving rules (Nature SR, 2025), or analyzing bias across CDIs (medRxiv, 2025). See also general PECARN data preprocessing ([clinical-rule-vetting](https://github.com/Yu-Group/rule-stress-testing) ) and ([clinical-self-verification](https://github.com/microsoft/clinical-self-verification) ). Also working on an LLM/EHR pipeline for this culminating in [HACHI](https://github.com/jjfenglab/HACHI), and previously [BC-LLM](https://github.com/jjfeng/bc-llm).
**[covid19-severity-prediction](https://github.com/Yu-Group/covid19-severity-prediction) ** Extensive COVID-19 data + forecasting for counties and hospitals ([HDSR 2021](https://hdsr.mitpress.mit.edu/pub/p6isyf0g/release/4))
**[molecular-partner-prediction](https://github.com/Yu-Group/molecular-partner-prediction) ** Predicting successful CME events using only clathrin markers
## Various aspects of deep learning and machine learning
**[gan-vae-pretrained-pytorch](https://github.com/csinva/gan-vae-pretrained-pytorch) ** Pretrained GANs + VAEs + classifiers for MNIST/CIFAR in pytorch
**[gpt-paper-title-generator](https://github.com/csinva/gpt-paper-title-generator) ** Generating paper titles with GPT-2
**[disentangled-attribution-curves](https://github.com/csinva/disentangled-attribution-curves) ** Attribution curves for interpreting tree ensembles trees ([arxiv 2019](https://arxiv.org/abs/1905.07631))
**[matching-with-gans](https://github.com/csinva/matching-with-gans) ** Matching in GAN latent space for better bias benchmarking. ([CVPR workshop 2021](https://arxiv.org/abs/2103.13455))
**[data-viz-utils](https://github.com/csinva/data-viz-utils) ** Functions for easily making publication-quality figures with matplotlib
**[mdl-complexity](https://github.com/csinva/mdl-complexity) ** Revisiting complexity and the bias-variance tradeoff ([JMLR 2021](https://arxiv.org/abs/2006.10189))
## Projects advised
**[pasta](https://github.com/QingruZhang/PASTA) ** Post-hoc Attention Steering for LLMs ([ICLR 2024](https://arxiv.org/abs/2311.02262)), led by [Qingru Zhang](https://github.com/QingruZhang)
**[meta-tree](https://github.com/EvanZhuang/MetaTree) ** Learning a Decision Tree Algorithm with Transformers ([TMLR 2024](https://arxiv.org/abs/2402.03774)), led by [Yufan Zhuang](https://github.com/EvanZhuang)
**[sim-dino](https://github.com/RobinWu218/SimDINO) ** Simplifying DINO via coding rate regularization ([ICML 2025](https://arxiv.org/abs/2502.10385)), led by [Ziyang Wu](https://github.com/RobinWu218)
**[explanation-consistency-finetuning](https://github.com/yandachen/explanation-consistency-finetuning) ** Consistent Natural-Language Explanations ([COLING 2025](https://arxiv.org/abs/2401.13986)), led by [Yanda Chen](https://github.com/yandachen)
**[induction-gram](https://github.com/ejkim47/induction-gram) ** Interpretable Language Modeling via Induction-head Ngram Models ([arXiv 2024](https://arxiv.org/abs/2401.13986)), led by [Eunji Kim](https://github.com/ejkim47) & [Sriya Mantena](https://github.com/SriyaM)
## Open-source contributions
Major: **[autogluon](https://github.com/awslabs/autogluon) , [big-bench](https://github.com/google/BIG-bench) , [nl-augmenter](https://github.com/GEM-benchmark/NL-Augmenter) **
Minor: **[conference-acceptance-rates](https://github.com/lixin4ever/Conference-Acceptance-Rate) , [iterative-random-forest](https://github.com/Yu-Group/iterative-Random-Forest) , [interpretable-ml-book](https://github.com/christophM/interpretable-ml-book) , [awesome-interpretable-machine-learning](https://github.com/lopusz/awesome-interpretable-machine-learning) , [awesome-machine-learning-interpretability](https://github.com/jphall663/awesome-machine-learning-interpretability) , [awesome-llm-interpretability](https://github.com/JShollaj/awesome-llm-interpretability) , [executable-books](https://github.com/executablebooks/meta) , [deep-fMRI-dataset](https://github.com/HuthLab/deep-fMRI-dataset) **
## Mini-projects
**[hummingbird-tracking](https://github.com/csinva/hummingbird-tracking), [imodels-experiments](https://github.com/Yu-Group/imodels-experiments), [cookiecutter-ml-research](https://github.com/csinva/cookiecutter-ml-research), [nano-descriptions](https://github.com/csinva/nano-descriptions), [news-title-bias](https://github.com/csinva/news-title-bias), [java-mini-games](https://github.com/csinva/mini-games), [imodels-data](https://github.com/csinva/imodels-data), [news-balancer](https://github.com/csinva/news-balancer), [arxiv-copier](https://github.com/csinva/arxiv-copier), [dnn-experiments](https://github.com/csinva/dnn-experiments), [max-activation-interpretation-pytorch](https://github.com/csinva/max-activation-interpretation-pytorch), [acronym-generator](https://github.com/csinva/acronym-generator), [hpa-interp](https://github.com/csinva/hpa-interp), [sensible-local-interpretations](https://github.com/csinva/sensible-local-interpretations), [global-sports-analysis](https://github.com/csinva/global-sports-analysis), [mouse-brain-decoding](https://github.com/csinva/mouse-brain-decoding), ...**