Tag
The paper introduces MIL-BERT, an algorithm for classifying arbitrarily long texts by selecting relevant excerpts, achieving state-of-the-art results on multiple datasets with performance and explanatory guarantees.
This paper proposes BridgeMIL, a two-stage framework for EEG-based disease diagnosis that decouples instance representation learning from subject-level supervision using multiple instance learning. It achieves state-of-the-art accuracy on three EEG disease datasets, outperforming strong baselines.
This paper presents an Attention-Based Multiple Instance Learning (ABMIL) framework that leverages patient-level labels from cancer registries to train deep learning classifiers for tumor group classification without requiring per-report annotations, achieving a macro F1 of 0.83 on tasks at the BC Cancer Registry.
Introduces a distribution-based multiple instance learning framework using zero-inflated beta modeling to improve tumor proportion score prediction in non-small cell lung cancer from histopathology slides.
This paper proposes Outcome-Supervised Process Reward Modeling via Learnable Credit Assignment (LCA), a framework that jointly learns credit assignment and reward modeling under a weakest-link principle, formulated as a Multiple Instance Learning problem with Softmax-Weighted-Sum pooling. Experiments show it outperforms existing outcome-supervised PRMs across multiple tasks.
This paper introduces QG-MIL, a gated transformer aggregator that mitigates attention concentration in multiple instance learning for medical imaging, achieving domain-agnostic performance without auxiliary losses.
This paper proposes a Perceiver-style architecture pretrained on synthetic bag-structured data to enable efficient, task-adaptive classification from few labeled examples in multiple instance learning, outperforming supervised baselines across twelve benchmarks.
This paper analyzes hallucination detection in LLMs, proposing a max-pooling approach that improves efficiency by eliminating costly semantic consistency computations while maintaining competitive performance.
This paper proposes a novel framework combining Large Language Models with Multiple-Instance Learning to detect cognitive distortions in mental health texts by decomposing utterances into Emotion, Logic, and Behavior components and using multi-view gated attention for classification. The approach demonstrates improved performance on Korean and English datasets, particularly for distortions with high interpretive ambiguity.