Search papers, labs, and topics across Lattice.
The paper introduces HY-WU, a novel memory-first adaptation framework designed to address the challenges of continual learning and personalization in foundation models deployed in dynamic environments. HY-WU uses a neural module to generate instance-specific weight updates on-the-fly, effectively implementing functional memory at the operator level. This approach allows for adaptation without overwriting shared parameters, mitigating the risks of catastrophic forgetting and interference.
Stop retraining your foundation model for every new task: HY-WU generates instance-specific weight updates on-the-fly, enabling continual learning without catastrophic forgetting.
Foundation models are transitioning from offline predictors to deployed systems expected to operate over long time horizons. In real deployments, objectives are not fixed: domains drift, user preferences evolve, and new tasks appear after the model has shipped. This elevates continual learning and instant personalization from optional features to core architectural requirements. Yet most adaptation pipelines still follow a static weight paradigm: after training (or after any adaptation step), inference executes a single parameter vector regardless of user intent, domain, or instance-specific constraints. This treats the trained or adapted model as a single point in parameter space. In heterogeneous and continually evolving regimes, distinct objectives can induce separated feasible regions over parameters, forcing any single shared update into compromise, interference, or overspecialization. As a result, continual learning and personalization are often implemented as repeated overwriting of shared weights, risking degradation of previously learned behaviors. We propose HY-WU (Weight Unleashing), a memory-first adaptation framework that shifts adaptation pressure away from overwriting a single shared parameter point. HY-WU implements functional (operator-level) memory as a neural module: a generator that synthesizes weight updates on-the-fly from the instance condition, yielding instance-specific operators without test-time optimization.
