technique
Variational Information Bottleneck
techniqueactiveprovisional
variational-information-bottleneck-c8a0f8d6·1 events·first seen 20d agoAliases: Variational Information Bottleneck
Co-occurring entities
More like this (12)
amortized variational inferenceVariational Autoencoder (VAE)ELBO variance minimizationKullback-Leibler divergenceIntervention-Aware Variational Quantum Differentiable Predictive ControlVector QuantizationIn-context Vector Distillationmutual informationChannel-wise Vector QuantizationConcept Vector ExtractionJanus: A Benchmark for Goal-Conditioned Information Distortion in LLMsvisual language model
Recent events (1)
AREA: Attribute Extraction and Aggregation for CLIP-Based Class-Incremental Learning
AREA is a new method for CLIP-based Class-Incremental Learning (CIL) that decomposes the classification process into attribute extraction and aggregation stages to combat catastrophic forgetting. Extraction is stabilized by anchoring visual and textual attributes on a hyperspherical embedding space via principal geodesic analysis, while aggregation uses lightweight task-specific experts regularized by a variational information bottleneck. Inference employs optimal transport routing over task attribute manifolds. The method is reported to consistently outperform state-of-the-art CIL approaches and is accepted at ICML 2026.