Interleaved Learning

Interleaved learning, a biologically-inspired training approach, focuses on improving machine learning model performance by strategically alternating between different tasks or data streams during training. Current research explores its application across diverse areas, including reinforcement learning, vision-language models, and anomaly detection, often employing techniques like contrastive learning, concept bottleneck models, and adaptive thresholding to enhance efficiency and generalization. This approach holds significant promise for creating more robust and adaptable AI systems, particularly in scenarios with limited labeled data or evolving data distributions, and is driving advancements in both theoretical understanding and practical applications.

Papers

November 4, 2024

PPLLaVA: Varied Video Sequence Understanding With Prompt Guidance
Ruyang Liu, Haoran Tang, Haibo Liu, Yixiao Ge, Ying Shan, Chen Li, Jiankun Yang
Video Understanding Video LLM Long Video Understanding Procedural Activity Interleaved Learning Prompt Alignment Prompt Pool

July 22, 2024

Concept-Based Interpretable Reinforcement Learning with Limited to No Human Labels
Zhuorui Ye, Stephanie Milani, Geoffrey J. Gordon, Fei Fang
Reinforcement Learning Concept Bottleneck Model Concept Based Explanation Human Label Interleaved Learning Limited Number

June 11, 2024

Vision Model Pre-training on Interleaved Image-Text Data via Latent Compression Learning
Chenyu Yang, Xizhou Zhu, Jinguo Zhu, Weijie Su, Junjie Wang, Xuan Dong, Wenhai Wang, Lewei Lu, Bin Li, Jie Zhou, Yu Qiao, Jifeng Dai
Pre Trained Vision Model Compressed Representation Interleaved Learning Interleaved Image Text

January 24, 2024

Interleaving One-Class and Weakly-Supervised Models with Adaptive Thresholding for Unsupervised Video Anomaly Detection
Yongwei Nie, Hao Huang, Chengjiang Long, Qing Zhang, Pradipta Maji, Hongmin Cai
Full Model Video Anomaly Detection One Class Classification Automatic Thresholding Adaptive Thresholding Interleaved Learning

January 1, 2024

COSMO: COntrastive Streamlined MultimOdal Model with Interleaved Pre-Training
Alex Jinpeng Wang, Linjie Li, Kevin Qinghong Lin, Jianfeng Wang, Kevin Lin, Zhengyuan Yang, Lijuan Wang, Mike Zheng Shou
Contrastive Learning Contrastive Loss Video Text Video Text Task Interleaved Learning Autoregressive Vision Language Model

December 12, 2023

Interfacing Foundation Models' Embeddings
Xueyan Zou, Linjie Li, Jianfeng Wang, Jianwei Yang, Mingyu Ding, Zhengyuan Yang, Feng Li, Hao Zhang, Shilong Liu, Arul Aravinthan, Yong Jae Lee, Lijuan Wang
Foundation Model Multi Task Jina Embeddings COCO Dataset Lightweight Transformer Interleaved Learning

September 22, 2023

A Simple Illustration of Interleaved Learning using Kalman Filter for Linear Least Squares
Majnu John, Yihren Wu
Machine Learning Algorithm Least Square Kalman Filter Linear Ordered Data Optimization Framework Interleaved Learning

December 2, 2022

On the Change of Decision Boundaries and Loss in Learning with Concept Drift
Fabian Hinder, Valerie Vaquet, Johannes Brinkrolf, Barbara Hammer
LeArning Abstract Generalization Error Concept Drift Decision Boundary Pre Change Information Drift Detection Interest Loss Real Drift Interleaved Learning

July 7, 2022

Style Interleaved Learning for Generalizable Person Re-identification
Wentao Tan, Changxing Ding, Pengfei Wang, Mingming Gong, Kui Jia
Person Re Identification Backward Propagation Interleaving Method Generalizable Person Re Identification Interleaved Learning

March 22, 2022

Modelling continual learning in humans with Hebbian context gating and exponentially decaying task signals
Timo Flesch, David G. Nagy, Andrew Saxe, Christopher Summerfield
Continual LEArning Real Human Standard Deep Hebbian Learning Behavioral Signal Prefrontal Cortex Interleaved Learning Context Gating Hebbian Model

Interleaved Learning

Papers

PPLLaVA: Varied Video Sequence Understanding With Prompt Guidance

Concept-Based Interpretable Reinforcement Learning with Limited to No Human Labels

Vision Model Pre-training on Interleaved Image-Text Data via Latent Compression Learning

Interleaving One-Class and Weakly-Supervised Models with Adaptive Thresholding for Unsupervised Video Anomaly Detection

COSMO: COntrastive Streamlined MultimOdal Model with Interleaved Pre-Training

Interfacing Foundation Models' Embeddings

A Simple Illustration of Interleaved Learning using Kalman Filter for Linear Least Squares

On the Change of Decision Boundaries and Loss in Learning with Concept Drift

Style Interleaved Learning for Generalizable Person Re-identification

Modelling continual learning in humans with Hebbian context gating and exponentially decaying task signals