Search papers, labs, and topics across Lattice.
Fuzhou University
5
0
9
LLMs can now dynamically re-attend to procedural rules, boosting their reasoning accuracy by up to 19 points in complex scenarios.
MGSI reveals that integrating audio and visual sentiment cues at multiple temporal scales can dramatically enhance sentiment analysis accuracy in LLMs.
Hallucinations in sentimental image captions can be drastically reduced by grounding sentiment in verifiable object-level evidence.
Get robust and interpretable models by combining adversarial training with explanation-guided learning, achieving a whopping +37% boost in adversarial accuracy on OOD data.
Get better few-shot graph learning by explaining your predictions: BAED extracts informative subgraphs to improve accuracy and interpretability.
