Code for the paper "FOCIL: Finetune-and-Freeze for Online Class-Incremental Learning by Training Randomly Pruned Sparse Experts"

This project is to showcase how to use Knowledge distillation technique to develop incremental learning system.This is a implementation of the paper https://arxiv.org/abs/1811.08051.

[ECCV 2024] Official Implementation of An Incremental Unified Framework for Small Defect Inspection

[ECMLPKDD 2022] "Avoiding Forgetting and Allowing Forward Transfer in Continual Learning via Sparse Networks" by by Ghada Sokar, Decebal Constantin Mocanu, and Mykola Pechenizkiy.

Repo for competition track Lifelong Robotic Vision, IROS 2019.

Model Retraining & Incremental Learning in Recommender Systems

Incremental Learning with Policy Gradient Methods and Eligibility Traces

Modeling student's knowledge over time using Riiid's EdNet dataset comprising of 100M+ student interactions. EDA and Feature Engineering was performed on the data and models were trained on important features.

The project leverages a pre-trained teacher model (SimpleCNN) to guide the training of a student model (CTNet). The training process combines cross-entropy loss for true labels with KL divergence loss for soft targets provided by the teacher model.

Welcome to my GitHub page!

🤖[MICCAI 2023] The official repository for paper "L3DMC: Lifelong Learning using Distillation via Mixed-Curvature Space"

Enhance SVM and incremental SVM machine learning models for road accident severity prediction

Data Efficient Incremental Learning Via Attentive Knowledge Replay (IEEE SMC 2023)

How close a computer is from being attacked by malware (Kaggle)

detecting domain boundaries during inference

Cost-Free Incremental Learning

Contains the examples which covers how to incrementally train, how to implement learning_rate scheduler, and how to implement custom objective and evaluation function in case of lightgbm/xgboost models.

Reproduction of existing baselines that address incremental learning issues and definition of new approaches to overcome existing limitations

Developed an interactive engine that is capable of visualizing the active(incremental) mapping of a bot along with its sensor interactions with its environment.

An anthology of recent continual learning papers, where people interested in this fascinating topic can start discovering its multidimensional representations.