Goto

Collaborating Authors

 Industry




FormalizingtheGeneralization-ForgettingTrade-Off inContinualLearning

Neural Information Processing Systems

In continual learning (CL), we incrementally adapt a model to learn tasks (defined according to the problem at hand) observed sequentially. CL has two main objectives: maintain long-term memory (remember previous tasks) and navigate new experiences continually (quickly adapt to newtasks).






ConMe: RethinkingEvaluationofCompositional ReasoningforModernVLMs-SupplementaryMaterial-AnonymousAuthor(s) Affiliation Address email

Neural Information Processing Systems

As an example, for an image taken on the ground, two text options24 are: {Several vehicles providing ground transportation are shown in25 the photo: streetcar, tour bus, classic car, and family cars.}