Task-agnostic Continual Learning with Hybrid Probabilistic Models

2021/06/24 by Polina Kirichenko, Kirichenko, Polina, Mehrdad Farajtabar +15 · 1 citation
Computer Science · #Adversarial Robustness in Machine Learning #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning and ELM

paper · pdf · doi:10.48550/arxiv.2106.12772

openalex publication_date 2021/06/24 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Learning new tasks continuously without forgetting on a constantly changing data distribution is essential for real-world problems but extremely challenging for modern deep learning. In this work we propose HCL, a Hybrid generative-discriminative approach to Continual Learning for classification. We model the distribution of each task and each class with a normalizing flow. The flow is used to learn the data distribution, perform classification, identify task changes, and avoid forgetting, all leveraging the invertibility and exact likelihood which are uniquely enabled by the normalizing flow model. We use the generative capabilities of the flow to avoid catastrophic forgetting through generative replay and a novel functional regularization technique. For task identification, we use state-of-the-art anomaly detection techniques based on measuring the typicality of the model's statistics. We demonstrate the strong performance of HCL on a range of continual learning benchmarks such as split-MNIST, split-CIFAR, and SVHN-MNIST.

Citations

• Adam: A Method for Stochastic Optimization
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Revisiting Natural Gradient for Deep Networks
• Semi-Supervised Learning with Deep Generative Models
• Gradient-based learning applied to document recognition
• Functional Regularisation for Continual Learning with Gaussian Processes
• Semi-Conditional Normalizing Flows for Semi-Supervised Learning
• Learning to Learn without Forgetting by Maximizing Transfer and Minimizing Interference
• FFJORD: Free-form Continuous Dynamics for Scalable Reversible Generative Models
• Continual Unsupervised Representation Learning
• Variational Prototype Replays for Continual Learning
• Hybrid Models for Open Set Recognition
• Why Normalizing Flows Fail to Detect Out-of-Distribution Data
• Density of States Estimation for Out-of-Distribution Detection
• Pseudo-Rehearsal for Continual Learning with Normalizing Flows
• Learning latent representations across multiple data domains using Lifelong VAEGAN
• Learning Multiple Layers of Features from Tiny Images

Cited by

• Deep Generative Continual Learning using Functional LoRA: FunLoRA

Related

• Generative Kernel Continual learning
• Continual Classification Learning Using Generative Models
• FoCL: Feature-Oriented Continual Learning for Generative Models
• Continual Learning with Fully Probabilistic Models
• Hierarchically Structured Task-Agnostic Continual Learning