Optical Coherence Tomography (OCT) is a widely adopted imaging technique in ophthalmology, yet its clinical interpretation remains prone to inconsistency and subjectivity. This study presents a retrieval-augmented generation (RAG) framework aimed at improving both diagnostic accuracy and interpretability in OCT-based analysis. The pipeline integrates a dual-headed convolutional neural network with a case-based retrieval system and a language model capable of generating structured clinical summaries. A ResNet-50 model was trained to classify diabetic retinopathy (DR), diabetic macular edema (DME), and key retinal biomarkers using a public OCT dataset. Feature embeddings generated by the model were indexed into a similarity-search database to enable retrieval of clinically relevant cases. Retrieved scans, along with classification outputs, were used to prompt a large language model to generate concise diagnostic reports that reflect both the model’s predictions and visually aligned precedents. The system achieved 89% accuracy in distinguishing DR from DME in a hold-out test set. Interpretability was also demonstrated through consistent retrieval of semantically similar cases. Generated reports were found to be readable and aligned with clinical reasoning. By combining classification performance with transparent, case-grounded explanations, the RAG approach enhances clinical confidence and supports decision-making. This work highlights the potential of retrieval-augmented pipelines to reduce variability in OCT interpretation and to serve as a scalable, explainable foundation for future AI-driven diagnostic support in ophthalmology.

Improving Optical Coherence Tomography Interpretation With Retrieval-Guided Diagnosis and Text Generation

Chronic kidney disease (CKD) is a progressive and irreversible loss of kidney function that affected about 753 million people globally in 2016. There are often no noticeable symptoms until it reaches later stages which makes early identification important for prevention. Diabetes mellitus is among the leading causes of chronic kidney disease worldwide, with an estimated 40\% of diabetic patients developing diabetic kidney disease. This project applies machine learning techniques to predict self-reported kidney disease risk in older diabetic individuals (ages 55+) using data from the U.S. National Health and Nutrition Examination Survey (NHANES). I used survey- based variables such as blood pressure, glycohemoglobin, body mass index (BMI), physical activity, and demographics to create a classification model. After preprocessing and engineering features related to age, blood pressure, and activity levels, I trained logistic regression and random forest classifiers to identify individuals at higher risk. The random forest model achieved an area under the receiver operating characteristic (ROC) curve (AUC) of 0.93, while the logistic regression model achieved an ROC AUC of 0.87. These results show that both models can effectively identify high-risk individuals. Feature importance analysis identified blood pressure, age, BMI, and diabetes status as leading predictors. This study demonstrates that machine learning can be used to improve early detection of CKD risk in aging diabetic populations.

Early Detection of Kidney Disease in Aging Diabetic Populations Using Machine Learning

The global rise in the elderly population is putting an increasing strain on healthcare systems, particularly due to the increased chance of chronic illness, polypharmacy, and healthcare utilization. The traditional one-size-fits-all treatment approach often fails to address the unique and evolving needs of older adults, leading to adverse drug reactions, diminished quality of life, and inefficient use of medical resources. This study investigated the potential of personalized medicine — augmented by artificial intelligence (AI) — to improve elderly care through data-driven, individualized treatment plans. By integrating data from electronic health records, genomics, wearable devices, and lifestyle factors, we developed AI-enhanced models capable of tailoring treatments to each patient’s specific profile. These models allow for real-time treatment adjustments, predictive risk assessments, and optimized drug regimens. Our findings suggest that AI-driven personalized care can reduce adverse outcomes, enhance clinical effectiveness, and support proactive, scalable, and patient-centered healthcare for aging populations.

Improving Prediction of ADR in Elderly Patients Using Machine Learning Algorithms

Multi-modality pre-training paradigm that aligns protein sequences and biological descriptions has learned general protein representations and achieved promising performance in various downstream applications. However, these works were still unable to replicate the extraordinary success of language-supervised visual foundation models due to the ineffective usage of aligned protein-text paired data and the lack of an effective function-informed pre-training paradigm. To address these issues, this paper curates a large-scale protein-text paired dataset called ProtAnno with a property-driven sampling strategy, and introduces a novel function-informed protein pre-training paradigm. Specifically, the sampling strategy determines sampling probability based on the confidence and coverage of property annotations, balancing the data quality and data quantity in face of large-scale noisy data. Furthermore, motivated by significance of the protein specific functional mechanism, the proposed paradigm explicitly model protein static and dynamic functional segments by two segment-wise pre-training objectives, injecting fine-grained information in a function-informed manner. Leveraging all these innovations, we develop ProtCLIP, a multi-modality foundational model that comprehensively represents function-aware protein embeddings. On 22 different protein benchmarks within 5 types, including protein functionality classification, mutation effect prediction, cross-modality transformation, and protein-protein interaction prediction, our ProtCLIP consistently achieves the state-of-the-art performance, with remarkable improvements of 75% on average in five cross-modality transformation tasks, 59.9% in GO-CC and 39.7% in GO-BP protein function prediction. The experimental results demonstrate the extraordinary potential of ProtCLIP serving as the protein multi-modality foundation model.

ProtCLIP: Function-Informed Protein Multi-Modal Learning

Multi-objective evolutionary algorithms (MOEAs) are widely used for searching optimal solutions in complex multi-component applications. Traditional MOEAs for multi-component deep learning (MCDL) systems face challenges in enhancing the search efficiency while maintaining the diversity. To combat these, this paper proposes $\mu$MOEA, the first LLM-empowered adaptive evolutionary search methodology to detect safety violations in MCDL systems. Inspired by the context-understanding ability of Large Language Models (LLMs), $\mu$MOEA promotes the LLM to comprehend the optimization problem and generate an initial population tailed to evolutionary objectives. Subsequently, it employs adaptive selection and variation techniques to iteratively produce offspring, balancing the evolutionary efficiency and diversity.
During the evolutionary process, to navigate away from the local optima,  $\mu$MOEA integrates the evolutionary experience back into the LLM. This utilization harnesses the LLM's quantitative reasoning prowess to generate differential seeds, breaking away from current optimal solutions. We evaluate the effectiveness of \name in finding safety violations of MCDL systems, and compare its performance with state-of-the-art MOEA methods. Experimental results show that \name can significantly improve the efficiency and diversity of the multi-objective evolutionary search.

An LLM-Empowered Adaptive Evolutionary Algorithm for Multi-Component Deep Learning Systems

Considering the difficulty of interpreting generative model output, there is significant current research focused on determining meaningful evaluation metrics. Several recent approaches utilize "precision" and "recall," borrowed from the classification domain, to individually quantify the output fidelity (realism) and output diversity (representation of the real data variation), respectively. With the increase in metric proposals, there is a need for a unifying perspective, allowing for easier comparison and clearer explanation of their benefits and drawbacks. To this end, we unify a class of kth-nearest-neighbors (kNN)-based metrics under an information-theoretic lens using approaches from kNN density estimation. Additionally, we propose a tri-dimensional metric composed of Precision Cross-Entropy (PCE), Recall Cross-Entropy (RCE), and Recall Entropy (RE), which separately measure fidelity and two distinct aspects of diversity, inter- and intra-class. Our domain-agnostic metric, derived from the information-theoretic concepts of entropy and cross-entropy, can be dissected for both sample- and mode-level analysis. Our detailed experimental results demonstrate the sensitivity of our metric components to their respective qualities and reveal undesirable behaviors of other metrics.

A Unifying Information-theoretic Perspective on Evaluating Generative Models

The applicability of drug molecules across various clinical scenarios is significantly influenced by a diverse range of molecular properties. By leveraging the self-supervised conditions such as atom attributes and inter-atomic bonds, existing advanced molecular foundation models can generate expressive representations of these molecules. However, such models often overlook the fixed association patterns within molecules that influence physiological or chemical properties. In this paper, we introduce a novel association pattern-aware message passing method, which can serve as an effective yet general plug-and-play plugin, thereby enhancing the atom representations generated by molecular foundation models without requiring additional pretraining. Additionally, the molecular property-specific pattern libraries are constructed to collect the generated interpretable common patterns that bind to these properties. Extensive experiments conducted on 11 benchmark molecular property prediction tasks across 8 advanced molecular foundation models demonstrate significant superiority of the proposed method, with the performance improvements of up to approximately 20%. Furthermore, we specifically tailor a property-specific pattern library for blood-brain barrier penetration, which has undergone corresponding mechanistic validation. Our code and technical appendix are included in the Supplementary Materials and will be made publicly available in the future.

Association Pattern-enhanced Molecular Representation Learning

Significant efforts have been dedicated to integrating the powerful Large Language Models (LLMs) with diverse modalities, particularly focusing on the fusion of language, vision and audio data. However, the graph-structured data, which is inherently rich in structural and domain-specific knowledge, has not yet been gracefully adapted to LLMs. Existing methods either describe the graph with raw text, suffering the loss of graph structural information, or feed Graph Neural Network (GNN) embeddings into LLMs at the cost of losing explainable prompt semantics. To bridge this gap, we introduce an end-to-end modality-aligning framework for LLM-graph alignment: Dual-Residual Vector Quantized-Variational AutoEncoder, namely Dr.E. Our approach is purposefully designed to facilitate token-level alignment with LLMs, enabling an effective translation of the intrinsic `language’ of graphs into comprehensible natural language. We also manage to enhance LLMs’ more robust structural understanding of graphs by incorporating multiple views of the central nodes based on their surrounding nodes at various distances. Our experimental evaluations on standard graph tasks demonstrate competitive performance against other state-of-the-art (SOTA) approaches. Additionally, our framework ensures certain visual interpretability, efficiency, and robustness, marking the promising successful endeavor to achieve token-level alignment between LLMs and GNNs. Our code is available at: https://anonymous.4open.science/r/dre-817.

Multi-View Empowered Structural Graph Wordification for Language Models

Autonomous underwater vehicle (AUV) is crucial for marine applications such as ocean data collection, pollution monitoring, and navigation. However, their limited energy resources constrain their operational duration, posing a significant challenge for long-term operations. Due to the complex and unpredictable nature of the underwater environment, AUVs allocate energy to their sensing systems to sense the surrounding environment and avoid obstacles. Existing methods focus on reducing energy consumption on AUV computing and movement, neglecting sensing energy consumption and few attempts have been made to balance the AUV energy and sensing ability with a flexible sensing system. Along these lines, we consider both AUV energy consumption and flexible sensing abilities, and propose a deep reinforcement learning-based method to Reduce Energy Consumption by AUV Sensing system (RECS). Specifically, we build an AUV sensing system in a 2-dimension space, with controllable 8-direction sensing abilities to collect the environment information dynamically. Then we divide the underwater environment into several areas and assign weights on the edges of areas based on the AUV planned path. Additionally, we dynamically switch the sensors in different directions and radii to sense the edges of the area where the AUV is located. The Artificial Potential Field (APF) method is employed to re-plan the AUV path to avoid obstacles and reach the target point effectively. Experimental results demonstrate that compared to full sensors on, our method reduces energy consumption by 53.48% and is capable of generalizing to varying environments and varying sensing system radii.

Reducing AUV Energy Consumption through Dynamic Sensor Directions Switching via Deep Reinforcement Learning

Point-of-Interest (POI) recommendation aims to predict users' future locations based on their historical check-ins. Despite the success of recent deep learning approaches in capturing POI semantics and user behavior, they continue to face the persistent problem of data sparsity and incompleteness. In this paper, we introduce Multi-Objective Adversarial Imitation Recommender (MOAIR), a novel framework that integrates Generative Adversarial Imitation Learning with multi-objective to address this issue. MOAIR effectively captures user behavior patterns and spatial-temporal contextual information via graph-enhanced self-supervised state encoder and overcomes data sparsity by robustly learning from limited data and generating diverse samples. By accommodating diverse user patterns in the training data, the framework also mitigates the typical mode-collapse issue in generative adversarial learning and thus enhances the overall performance. MOAIR employs a multi-objective imitation learning architecture where the imitation learning agent (IL agent) explores the POI space and receives multifaceted reward signals. Utilizing the Paralleled Proximal Policy Optimization (3PO) framework to optimize multi-objectives, the IL agent ensures efficient and stable policy updates. Additionally, to address the issue of high noise in POI recommendation scenarios, we use a novel generative way to define our policy net and incorporate a variational bottleneck for regularization to enhance the stability of adversarial learning. Comprehensive experiments reveal the superior performance for MOAIR compared to other baseline approaches, especially with sparse training data.

POI Recommendation via Multi-Objective Adversarial Imitation Learning

Traffic prediction is pivotal in intelligent transportation systems. Existing works mainly focus on improving overall accuracy, overlooking a crucial problem of whether prediction results will lead to biased decisions by transportation authorities. In practice, the uneven deployment of traffic sensors in different urban areas produces imbalanced data, making the traffic prediction model fail in some areas and leading to unfair regional decision-making that eventually severely affects residents' equity and quality of life. Additionally, existing fairness machine learning models fail to preserve fair traffic prediction for a prolonged time. Although they can achieve fairness at certain time points, such static fairness will be broken as the traffic conditions change. To fill this research gap, we investigate prolonged fair traffic prediction, introduce two novel fairness definitions tailored to dynamic traffic scenarios, and propose a prolonged fairness traffic prediction framework, namely FairTP.  We argue that fairness in traffic scenarios changes dynamically over time and across areas. The prediction accuracy of each traffic sensor or city area should take turns to "sacrifice" or "benefit" in order for the model to achieve fairness, eventually achieving prolonged fairness when such "sacrifice" and "benefit" are similar within a period of time. Accordingly, we first defined region-based static fairness and sensor-based dynamic fairness. Next, we designed a state identification module in FairTP to discriminate between states of "sacrifice" or "benefit" to enable prolonged fairness-aware traffic predictions. Lastly, a state-guided balanced sampling strategy is designed to select training examples to promote prediction fairness further, mitigating the performance disparities among regions with imbalanced traffic sensors. Extensive experiments in two real-world datasets show that FairTP significantly improves prediction fairness without causing much accuracy degradation.

FairTP: A Prolonged Fairness Framework for Traffic Prediction

Given an edge-colored graph, the goal of the Proportionally Fair Matching problem is to find a *maximum weight matching* while ensuring proportional representation of each color. The colors may correspond to demographic groups or other protected traits where we seek to ensure roughly equal representation from each group. It is known that, assuming ETH, it is impossible to approximate the problem with $\ell$ colors in time $2^{o(\ell)} n^{\mathcal{O}(1)}$ (i.e., subexponential in $\ell$). Further, even determining the existence of a non-empty matching satisfying proportionality is NP-Hard.
  
To overcome this hardness, we relax the stringent proportional fairness constraints to a probabilistic notion. We introduce a notion we call $\eta$-ProbablyApproximatelyFair, where we ensure proportionality to within a constant factor violation with high probability. We propose and analyze simple, fast algorithms with constant-factor approximation guarantees that satisfy $\eta$-ProbablyApproximateFairness for the Proportionally Fair Matching problem.

Underline digital video library

Improving Optical Coherence Tomography Interpretation With Retrieval-Guided Diagnosis and Text Generation

Early Detection of Kidney Disease in Aging Diabetic Populations Using Machine Learning

Improving Prediction of ADR in Elderly Patients Using Machine Learning Algorithms

ProtCLIP: Function-Informed Protein Multi-Modal Learning

An LLM-Empowered Adaptive Evolutionary Algorithm for Multi-Component Deep Learning Systems

A Unifying Information-theoretic Perspective on Evaluating Generative Models

Association Pattern-enhanced Molecular Representation Learning

Multi-View Empowered Structural Graph Wordification for Language Models

Reducing AUV Energy Consumption through Dynamic Sensor Directions Switching via Deep Reinforcement Learning

POI Recommendation via Multi-Objective Adversarial Imitation Learning

FairTP: A Prolonged Fairness Framework for Traffic Prediction

Proportionally Fair Matching via Randomized Rounding

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES