Overcoming Uncertain Incompleteness for Robust Multimodal Sequential Diagnosis Prediction via Curriculum Data Erasing Guided Distillation

2024-07-28

AI Summary

To address pervasive modality missingness and uncertainty in clinical visit sequences, this paper proposes NECHO v2—a robust framework for multimodal temporal diagnostic prediction. Methodologically, it introduces (1) a novel curriculum-based stochastic data erasure-guided knowledge distillation mechanism that dynamically emulates realistic missing patterns to enhance model resilience; and (2) a unified distillation strategy integrating modality-level contrastive learning, hierarchical representation modeling, and stochastic Transformer alignment to enable robust cross-modal and cross-granularity feature transfer. Evaluated on multiple balanced and imbalanced incomplete medical datasets, NECHO v2 consistently outperforms state-of-the-art baselines, achieving significant improvements in both diagnostic accuracy and prediction stability. The framework demonstrates superior generalizability under heterogeneous missingness patterns and varying data completeness levels, advancing the reliability of AI-assisted clinical decision support systems.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

In this paper, we present NECHO v2, a novel framework designed to enhance the predictive accuracy of multimodal sequential patient diagnoses under uncertain missing visit sequences, a common challenge in real clinical settings. Firstly, we modify NECHO, designed in a diagnosis code-centric fashion, to handle uncertain modality representation dominance under the imperfect data. Secondly, we develop a systematic knowledge distillation by employing the modified NECHO as both teacher and student. It encompasses a modality-wise contrastive and hierarchical distillation, transformer representation random distillation, along with other distillations to align representations between teacher and student tightly and effectively. We also propose curriculum learning guided random data erasing within sequences during both training and distillation of the teacher to lightly simulate scenario with missing visit information, thereby fostering effective knowledge transfer. As a result, NECHO v2 verifies itself by showing robust superiority in multimodal sequential diagnosis prediction under both balanced and imbalanced incomplete settings on multimodal healthcare data.

Problem

• Medical Diagnosis
• Uncertainty Handling
• Prediction Accuracy

Innovation

• NECHO v2
• Multi-modal Medical Information
• Predictive Accuracy under Incomplete Data