@NeurIPSConf @TrlWorkshop 6/6: In conclusion, RELOAD is an effective algorithm for unlearning arbitrary parts of the training set, and provides strong privacy guarantees for forgotten data.

@NeurIPSConf @TrlWorkshop 5/6: Using TabNet attention masks we show how RELOAD removes dependence of model inference on forgotten features.

@NeurIPSConf @TrlWorkshop 4/6: We conducted experiments on forgetting random samples and entire features from the training set, consistently outperforming unlearning baselines and protecting user privacy.

@NeurIPSConf @TrlWorkshop 3/6: Key Idea: We compare cached end-of-training gradients to those on the remaining data to identify parameters in the model to reset.

@NeurIPSConf @TrlWorkshop 2/6: Key Motivation: In unlearning, we typically require access to the set of data being forgotten. How can we unlearn, without that data?

RT @adibvafa: Canada's Largest AI Hackathon, @GenAIGenesis, will be back in Toronto, March 2025. Last year we had 1,200+ registrations fro…