REVIEW 4 cited by
The Effect of Task Ordering in Continual Learning
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
We investigate the effect of task ordering on continual learning performance. We conduct an extensive series of empirical experiments on synthetic and naturalistic datasets and show that reordering tasks significantly affects the amount of catastrophic forgetting. Connecting to the field of curriculum learning, we show that the effect of task ordering can be exploited to modify continual learning performance, and present a simple approach for doing so. Our method computes the distance between all pairs of tasks, where distance is defined as the source task curvature of a gradient step toward the target task. Using statistically rigorous methods and sound experimental design, we show that task ordering is an important aspect of continual learning that can be modified for improved performance.
Forward citations
Cited by 4 Pith papers
-
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.
-
PATH-Bench: Path-Dependent Evaluation of Lifelong Agents
A benchmark that controls task order to measure transfer, retention, and path-dependence in lifelong LLM agents, applied to code and tool-use tasks.
-
Optimal Task Order for Continual Learning of Multiple Tasks
In a linear teacher-student model, optimal continual learning orders place the least typical tasks first and make neighboring tasks dissimilar, and these rules transfer to image classification.
-
Sequence Transferability and Task Order Selection in Continual Learning
The paper proposes two sequence-level transferability scores, TFT and TRT, that correlate with continual learning accuracy, and a greedy task-order heuristic, HCTOS, that beats random ordering in a narrow set of experiments.
Discussion (0). Continue with ORCID to comment.