--Reliable multi-source fusion is crucial for robust perception in autonomous systems. However, evaluating fusion performance independently of detection errors remains challenging. This work introduces a systematic evaluation framework that injects controlled noise into ground-truth bounding boxes to isolate the fusion process. We then propose Unified Kalman Fusion (UniKF), a late-fusion algorithm based on Kalman filtering to merge Bird's Eye View (BEV) detections while handling synchronization issues. Experiments show that UniKF outperforms baseline methods across various noise levels, achieving up to 3 lower object's positioning and orientation errors and 2 lower dimension estimation errors, while maintaining near-perfect precision and recall between 99. 5% and 100%. Accurate perception is fundamental for autonomous driving, especially in complex urban settings where sensor occlusions, limited range, and adverse weather degrade detection quality [1]. Collaborative perception, enabled by onboard sensors' communication and V ehicle-to-Everything (V2X) communication, enhances perception by sharing sensor data across multiple sensors or agents [2], [3]. Early fusion methods require high bandwidth and strict time synchronization. Deep fusion demands access to proprietary models, which is impractical due to privacy and intellectual property restrictions. Late fusion, which operates at the object detection level, offers a scalable, bandwidth-efficient, and detector-model-agnostic alternative.

Domain adaptation has become a widely adopted approach in machine learning due to the high costs associated with labeling data. It is typically applied when access to a labeled source domain is available. However, in real-world scenarios, privacy concerns often restrict access to sensitive information, such as fingerprints, bank account details, and facial images. A promising solution to this issue is Source-Free Unsupervised Domain Adaptation (SFUDA), which enables domain adaptation without requiring access to labeled target domain data. Recent research demonstrates that SFUDA can effectively address domain discrepancies; however, two key challenges remain: (1) the low quality of prototype samples, and (2) the incorrect assignment of pseudo-labels. To tackle these challenges, we propose a method consisting of three main phases. In the first phase, we introduce a Reliable Sample Memory (RSM) module to improve the quality of prototypes by selecting more representative samples. In the second phase, we employ a Multi-View Contrastive Learning (MVCL) approach to enhance pseudo-label quality by leveraging multiple data augmentations. In the final phase, we apply a noisy label filtering technique to further refine the pseudo-labels. Our experiments on three benchmark datasets - VisDA 2017, Office-Home, and Office-31 - demonstrate that our method achieves approximately 2 percent and 6 percent improvements in classification accuracy over the second-best method and the average of 13 well-known state-of-the-art approaches, respectively.

The increasing proliferation of vending machines in public and commercial environments has placed a growing emphasis on operational efficiency and customer satisfaction. Traditional maintenance approaches either reactive or time-based preventive are limited in their ability to preempt machine failures, leading to unplanned downtimes and elevated service costs. This research presents a novel predictive maintenance framework tailored for vending machines by leveraging Internet of Things (IoT) sensors and machine learning (ML) algorithms. The proposed system continuously monitors machine components and operating conditions in real time and applies predictive models to forecast failures before they occur. This enables timely maintenance scheduling, minimizing downtime and extending machine lifespan. The framework was validated through simulated fault data and performance evaluation using classification algorithms. Results show a significant improvement in early fault detection and a reduction in redundant service interventions. The findings indicate that predictive maintenance systems, when integrated into vending infrastructure, can transform operational efficiency and service reliability.

MolProphecy is a human-in-the-loop (HITL) multi-modal framework designed to integrate chemists' domain knowledge into molecular property prediction models. While molecular pre-trained models have enabled significant gains in predictive accuracy, they often fail to capture the tacit, interpretive reasoning central to expert-driven molecular design. To address this, MolProphecy employs ChatGPT as a virtual chemist to simulate expert-level reasoning and decision-making. The generated chemist knowledge is embedded by the large language model (LLM) as a dedicated knowledge representation and then fused with graph-based molecular features through a gated cross-attention mechanism, enabling joint reasoning over human-derived and structural features. Evaluated on four benchmark datasets (FreeSolv, BACE, SIDER, and ClinTox), MolProphecy outperforms state-of-the-art (SOTA) models, achieving a 15.0 percent reduction in RMSE on FreeSolv and a 5.39 percent improvement in AUROC on BACE. Analysis reveals that chemist knowledge and structural features provide complementary contributions, improving both accuracy and interpretability. MolProphecy offers a practical and generalizable approach for collaborative drug discovery, with the flexibility to incorporate real chemist input in place of the current simulated proxy--without the need for model retraining. The implementation is publicly available at https://github.com/zhangruochi/MolProphecy.

Many e-learning platforms assert their ability or potential to improve students' self-regulated learning (SRL), however the cyclical and undirected nature of SRL theoretical models represent significant challenges for representation within contemporary machine learning frameworks. We apply SRL-informed features to trace data in order to advance modelling of students' SRL activities, to improve predictability and explainability regarding the causal effects of learning in an eLearning environment. We demonstrate that these features improve predictive accuracy and validate the value of further research into cyclic modelling techniques for SRL.

An algorithm that outputs predictions about the state of the world will almost always be designed with the implicit or explicit goal of outputting accurate predictions (i.e., predictions that are likely to be true). In addition, the rise of increasingly powerful predictive algorithms brought about by the recent revolution in artificial intelligence has led to an emphasis on building predictive algorithms that are fair, in the sense that their predictions do not systematically evince bias or bring about harm to certain individuals or groups. This state of affairs presents two conceptual challenges. First, the goals of accuracy and fairness can sometimes be in tension, and there are no obvious normative guidelines for managing the trade-offs between these two desiderata when they arise. Second, there are many distinct ways of measuring both the accuracy and fairness of a predictive algorithm; here too, there are no obvious guidelines on how to aggregate our preferences for predictive algorithms that satisfy disparate measures of fairness and accuracy to various extents. The goal of this paper is to address these challenges by arguing that there are good reasons for using a linear combination of accuracy and fairness metrics to measure the all-things-considered value of a predictive algorithm for agents who care about both accuracy and fairness. My argument depends crucially on a classic result in the preference aggregation literature due to Harsanyi. After making this formal argument, I apply my result to an analysis of accuracy-fairness trade-offs using the COMPAS dataset compiled by Angwin et al.

Vertical Federated Learning (VFL) is a distributed AI software deployment mechanism for cross-silo collaboration without accessing participants' data. However, existing VFL work lacks a mechanism to audit the execution correctness of the inference software of the data party. To address this problem, we design a Vertical Federated Inference Auditing (VeFIA) framework. VeFIA helps the task party to audit whether the data party's inference software is executed as expected during large-scale inference without leaking the data privacy of the data party or introducing additional latency to the inference system. The core of VeFIA is that the task party can use the inference results from a framework with Trusted Execution Environments (TEE) and the coordinator to validate the correctness of the data party's computation results. VeFIA guarantees that, as long as the abnormal inference exceeds 5.4%, the task party can detect execution anomalies in the inference software with a probability of 99.99%, without incurring any additional online inference latency. VeFIA's random sampling validation achieves 100% positive predictive value, negative predictive value, and true positive rate in detecting abnormal inference. To the best of our knowledge, this is the first paper to discuss the correctness of inference software execution in VFL.

Membership inference attacks (MIAs) pose a significant threat to the privacy of machine learning models and are widely used as tools for privacy assessment, auditing, and machine unlearning. While prior MIA research has primarily focused on performance metrics such as AUC, accuracy, and TPR@low FPR - either by developing new methods to enhance these metrics or using them to evaluate privacy solutions - we found that it overlooks the disparities among different attacks. These disparities, both between distinct attack methods and between multiple instantiations of the same method, have crucial implications for the reliability and completeness of MIAs as privacy evaluation tools. In this paper, we systematically investigate these disparities through a novel framework based on coverage and stability analysis. Extensive experiments reveal significant disparities in MIAs, their potential causes, and their broader implications for privacy evaluation. To address these challenges, we propose an ensemble framework with three distinct strategies to harness the strengths of state-of-the-art MIAs while accounting for their disparities. This framework not only enables the construction of more powerful attacks but also provides a more robust and comprehensive methodology for privacy evaluation.

Artificial intelligence systems, especially those using machine learning, are being deployed in domains from hiring to loan issuance in order to automate these complex decisions. Judging both the effectiveness and fairness of these AI systems, and their human decision making counterpart, is a complex and important topic studied across both computational and social sciences. Within machine learning, a common way to address bias in downstream classifiers is to resample the training data to offset disparities. For example, if hiring rates vary by some protected class, then one may equalize the rate within the training set to alleviate bias in the resulting classifier. While simple and seemingly effective, these methods have typically only been evaluated using data obtained through convenience samples, introducing selection bias and label bias into metrics. Within the social sciences, psychology, public health, and medicine, audit studies, in which fictitious ``testers'' (e.g., resumes, emails, patient actors) are sent to subjects (e.g., job openings, businesses, doctors) in randomized control trials, provide high quality data that support rigorous estimates of discrimination. In this paper, we investigate how data from audit studies can be used to improve our ability to both train and evaluate automated hiring algorithms. We find that such data reveals cases where the common fairness intervention method of equalizing base rates across classes appears to achieve parity using traditional measures, but in fact has roughly 10% disparity when measured appropriately. We additionally introduce interventions based on individual treatment effect estimation methods that further reduce algorithmic discrimination using this data.

This paper describes our system for the SciVQA 2025 Shared Task on Scientific Visual Question Answering. Our system employs an ensemble of two Multimodal Large Language Models and various few-shot example retrieval strategies. The model and few-shot setting are selected based on the figure and question type. We also select answers based on the models' confidence levels. On the blind test data, our system ranks third out of seven with an average F1 score of 85.12 across ROUGE-1, ROUGE-L, and BERTS. Our code is publicly available.