Well-conditioned encoder Jacobians, achieved via residual paths and LeakyReLU activations, improve trimodal contrastive learning retrieval and linear-probe performance across objectives and datasets.
Learnable latent embeddings for joint behavioural and neural analysis
5 Pith papers cite this work, alongside 338 external citations. Polarity classification is still indexing.
5
Pith papers citing it
338
external citations · Crossref
representative citing papers
DYSCO jointly recovers latent trajectories and governing equations from noisy observations via multi-view contrastive learning, with theoretical guarantees up to affine indeterminacy.
Reasoning fine-tuning reorganizes chain-of-thought into more differentiated latent switching states, and pruning with those states beats self-consistency in 11 of 12 settings.
The paper introduces a time-resolved neural encoder combining Whisper embeddings with recurrent temporal modeling and soft attention to predict ECoG responses, finding strongest alignment in intermediate layers and anatomically coherent phoneme organization in electrodes.
VLEM couples vision-language embeddings to separate where/what/when attractor networks and claims robust, interpretable episodic memory in a simulated agent.
citing papers explorer
-
Beyond Objective Expressivity: Geometry Preservation in Multimodal Contrastive Learning
Well-conditioned encoder Jacobians, achieved via residual paths and LeakyReLU activations, improve trimodal contrastive learning retrieval and linear-probe performance across objectives and datasets.
-
Extracting Governing Equations from Latent Dynamics via Multi-View Contrastive Learning
DYSCO jointly recovers latent trajectories and governing equations from noisy observations via multi-view contrastive learning, with theoretical guarantees up to affine indeterminacy.
-
Reasoning Fine-Tuning Induces Persistent Latent Policy States
Reasoning fine-tuning reorganizes chain-of-thought into more differentiated latent switching states, and pruning with those states beats self-consistency in 11 of 12 settings.
-
Mapping Whisper Representations to Human ECoG Responses with Interpretable Time-Resolved Neural Encoding
The paper introduces a time-resolved neural encoder combining Whisper embeddings with recurrent temporal modeling and soft attention to predict ECoG responses, finding strongest alignment in intermediate layers and anatomically coherent phoneme organization in electrodes.
-
Towards a Vision-Language Episodic Memory Framework: Large-scale Pretrained Model-Augmented Hippocampal Attractor Dynamics
VLEM couples vision-language embeddings to separate where/what/when attractor networks and claims robust, interpretable episodic memory in a simulated agent.