LIBERO is a new benchmark for lifelong robot learning that evaluates transfer of declarative, procedural, and mixed knowledge across 130 manipulation tasks with provided demonstration data.
hub
Lifelong Learning with Dynamically Expandable Networks
21 Pith papers cite this work. Polarity classification is still indexing.
21
Pith papers citing it
abstract
We propose a novel deep network architecture for lifelong learning which we refer to as Dynamically Expandable Network (DEN), that can dynamically decide its network capacity as it trains on a sequence of tasks, to learn a compact overlapping knowledge sharing structure among tasks. DEN is efficiently trained in an online manner by performing selective retraining, dynamically expands network capacity upon arrival of each task with only the necessary number of units, and effectively prevents semantic drift by splitting/duplicating units and timestamping them. We validate DEN on multiple public datasets under lifelong learning scenarios, on which it not only significantly outperforms existing lifelong learning methods for deep networks, but also achieves the same level of performance as the batch counterparts with substantially fewer number of parameters. Further, the obtained network fine-tuned on all tasks obtained significantly better performance over the batch models, which shows that it can be used to estimate the optimal network structure even when all tasks are available in the first place.
hub tools
citation-role summary
background 3
citation-polarity summary
roles
background 3representative citing papers
Introduces replay-based continual learning with sequential invariance alignment to learn domain-invariant representations, outperforming baselines on generalization to unseen domains across six datasets in vision, medicine, manufacturing, and ecology.
SPA unlocks patch-level features in CLIP for class-incremental learning via semantic-guided selection and optimal transport alignment with class descriptions, plus projectors and pseudo-feature replay to reduce forgetting.
NetTailor adapts CNN architecture for new tasks by assembling pre-trained universal blocks with task-specific layers, trained via activation mimicry and complexity penalties to match accuracy while reducing size for simpler tasks.
TeLAPA preserves behaviorally diverse policy neighborhoods in a shared latent space, improving MiniGrid continual RL transfer, revisit recovery, and retention over single-model preservation.
CVLC fuses calibrated vision prototypes with LLM-generated language prototypes and applies dual coalescent projection plus latent space reservation to enable few-shot adaptation across sequential domains, reporting up to 16% gains over prior methods.
TASER dynamically expands and orthogonality-constrains atomic skills then routes them with task-conditioned gating, outperforming baselines on the new 19-task HeteroCLBench benchmark for heterogeneous continual learning.
A continual few-shot adaptation method combining binary cross-entropy and supervised contrastive losses with replay achieves a good trade-off between fast adaptation to unseen synthetic fingerprint styles and retention of known styles.
A hypernetwork generates clock-augmented stable neural ODEs (sNODEs) for scalable continual learning from demonstration, achieving O(N) training time via stochastic regularization while outperforming baselines on LfD tasks up to 26 skills and 32 dimensions.
A blank-slate neural network grows via expansion, generalization, forgetting, and backpropagation for lifelong learning with claimed gains in accuracy, efficiency, and versatility.
NORACL grows neurons only when saturation signals fire, matching or beating static continual-learning models that know the future task budget while using fewer parameters.
MERS improves replay buffer selection in continual learning by integrating supervised and self-supervised embeddings via a graph-based approach, outperforming single-embedding baselines on CIFAR-100 and TinyImageNet in low-memory regimes.
DRDN reports 77.19% average accuracy on CIFAR100-B0 (10 steps) in from-scratch ViT class-incremental learning by routing MIM reconstruction through the backbone and using per-layer task token expansion.
STAR-IOD applies scale-decoupled topology alignment and K-Means-based pseudo-label refinement to reduce catastrophic forgetting in remote sensing incremental object detection, reporting 1.7% and 2.1% mAP gains on new DIOR-IOD and DOTA-IOD datasets.
Growth during training inserts new units into a specialized trajectory, making them forward-active but backward-starved with weaker gradients than existing units.
BRAIN uses bias-mitigation continual learning with a new de-bias contrastive loss and angular forgetting mitigation to achieve SOTA performance on vision-brain understanding benchmarks despite brain signal inconsistencies across sessions.
SwitchMT uses adaptive task-switching in deep spiking Q-networks with active dendrites to reduce task interference in multi-task RL, achieving competitive Atari scores without added network complexity.
HEDP uses energy regularization inspired by Helmholtz free energy plus hybrid energy-distance weighting in prompts to improve domain selection and achieve a 2.57% accuracy gain on benchmarks like CORe50 while mitigating catastrophic forgetting.
Proposes a domain incremental continual learning benchmark for ICU time series model transportability across US regions and evaluates data replay and EWC methods.
SOR-SNN employs Self-Organizing Regulation networks to reorganize a single SNN into sparse pathways, achieving better performance, energy efficiency, memory use, backward transfer, and self-repair on continual learning tasks including CIFAR100 and ImageNet.
Pseudo-rehearsal method with cGAN-generated old-concept samples, balanced online recall, and concept contrastive loss for class-incremental learning on MNIST, Fashion-MNIST and SVHN.
citing papers explorer
-
LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning
LIBERO is a new benchmark for lifelong robot learning that evaluates transfer of declarative, procedural, and mixed knowledge across 130 manipulation tasks with provided demonstration data.
-
Continual Learning of Domain-Invariant Representations
Introduces replay-based continual learning with sequential invariance alignment to learn domain-invariant representations, outperforming baselines on generalization to unseen domains across six datasets in vision, medicine, manufacturing, and ecology.
-
Unlocking Patch-Level Features for CLIP-Based Class-Incremental Learning
SPA unlocks patch-level features in CLIP for class-incremental learning via semantic-guided selection and optimal transport alignment with class descriptions, plus projectors and pseudo-feature replay to reduce forgetting.
-
NetTailor: Tuning the Architecture, Not Just the Weights
NetTailor adapts CNN architecture for new tasks by assembling pre-trained universal blocks with task-specific layers, trained via activation mimicry and complexity penalties to match accuracy while reducing size for simpler tasks.
-
Beyond Single-Model Optimization: Preserving Plasticity in Continual Reinforcement Learning
TeLAPA preserves behaviorally diverse policy neighborhoods in a shared latent space, improving MiniGrid continual RL transfer, revisit recovery, and retention over single-model preservation.
-
Few-Shot Domain Incremental Learning via Continual Vision-Language Consolidation
CVLC fuses calibrated vision prototypes with LLM-generated language prototypes and applies dual coalescent projection plus latent space reservation to enable few-shot adaptation across sequential domains, reporting up to 16% gains over prior methods.
-
Task-Differentiated Atomic Skill Expansion and Routing for Continual Learning Across Highly Heterogeneous Tasks
TASER dynamically expands and orthogonality-constrains atomic skills then routes them with task-conditioned gating, outperforming baselines on the new 19-task HeteroCLBench benchmark for heterogeneous continual learning.
-
Continual Few-shot Adaptation for Synthetic Fingerprint Detection
A continual few-shot adaptation method combining binary cross-entropy and supervised contrastive losses with replay achieves a good trade-off between fast adaptation to unseen synthetic fingerprint styles and retention of known styles.
-
Scalable and Efficient Continual Learning from Demonstration via a Hypernetwork-generated Stable Dynamics Model
A hypernetwork generates clock-augmented stable neural ODEs (sNODEs) for scalable continual learning from demonstration, achieving O(N) training time via stochastic regularization while outperforming baselines on LfD tasks up to 26 skills and 32 dimensions.
-
Lifelong Learning Starting From Zero
A blank-slate neural network grows via expansion, generalization, forgetting, and backpropagation for lifelong learning with claimed gains in accuracy, efficiency, and versatility.
-
NORACL: Neurogenesis for Oracle-free Resource-Adaptive Continual Learning
NORACL grows neurons only when saturation signals fire, matching or beating static continual-learning models that know the future task budget while using fewer parameters.
-
Leveraging Complementary Embeddings for Replay Selection in Continual Learning with Small Buffers
MERS improves replay buffer selection in continual learning by integrating supervised and self-supervised embeddings via a graph-based approach, outperforming single-embedding baselines on CIFAR-100 and TinyImageNet in low-memory regimes.
-
DRDN: Decoupled Representation Dynamic Network for From-Scratch ViT Class-Incremental Learning
DRDN reports 77.19% average accuracy on CIFAR100-B0 (10 steps) in from-scratch ViT class-incremental learning by routing MIM reconstruction through the backbone and using per-layer task token expansion.
-
STAR-IOD: Scale-decoupled Topology Alignment with Pseudo-label Refinement for Remote Sensing Incremental Object Detection
STAR-IOD applies scale-decoupled topology alignment and K-Means-based pseudo-label refinement to reduce catastrophic forgetting in remote sensing incremental object detection, reporting 1.7% and 2.1% mAP gains on new DIOR-IOD and DOTA-IOD datasets.
-
On the Stability of Growth in Structural Plasticity
Growth during training inserts new units into a specialized trajectory, making them forward-active but backward-starved with weaker gradients than existing units.
-
BRAIN: Bias-Mitigation Continual Learning Approach to Vision-Brain Understanding
BRAIN uses bias-mitigation continual learning with a new de-bias contrastive loss and angular forgetting mitigation to achieve SOTA performance on vision-brain understanding benchmarks despite brain signal inconsistencies across sessions.
-
Scalable Multi-Task Learning through Spiking Neural Networks with Adaptive Task-Switching Policy for Intelligent Autonomous Agents
SwitchMT uses adaptive task-switching in deep spiking Q-networks with active dendrites to reduce task interference in multi-task RL, achieving competitive Atari scores without added network complexity.
-
HEDP: A Hybrid Energy-Distance Prompt-based Framework for Domain Incremental Learning
HEDP uses energy regularization inspired by Helmholtz free energy plus hybrid energy-distance weighting in prompts to improve domain selection and achieve a 2.57% accuracy gain on benchmarks like CORe50 while mitigating catastrophic forgetting.
-
A Domain Incremental Continual Learning Benchmark for ICU Time Series Model Transportability
Proposes a domain incremental continual learning benchmark for ICU time series model transportability across US regions and evaluates data replay and EWC methods.
-
Adaptive Reorganization of Neural Pathways for Continual Learning with Spiking Neural Networks
SOR-SNN employs Self-Organizing Regulation networks to reorganize a single SNN into sparse pathways, achieving better performance, energy efficiency, memory use, backward transfer, and self-repair on continual learning tasks including CIFAR100 and ImageNet.
-
Incremental Concept Learning via Online Generative Memory Recall
Pseudo-rehearsal method with cGAN-generated old-concept samples, balanced online recall, and concept contrastive loss for class-incremental learning on MNIST, Fashion-MNIST and SVHN.