Grading programming assignments is crucial for guiding students to improve their skills and coding styles. This study presents an automated grading framework, CodEv, which leverages Large Language Models (LLMs) provide consistent constructive feedback. We incorporate Chain of Thought (CoT) prompting techniques enhance the reasoning capabilities LLMs ensure that aligned with human evaluation. Our framework also integrates LLM ensembles accuracy consistency scores, along agreement tests...

Abstract Objective Machine learning (ML) algorithms are promising tools for managing anemia in hemodialysis (HD) patients. However, their efficacy predicting erythropoiesis-stimulating agents (ESAs) doses remains uncertain. This study aimed to evaluate the effectiveness of a contemporary artificial intelligence (AI) model prescribing ESA compared physicians HD Materials and Methods double-blinded control trial randomized participants into traditional doctor (Dr) AI groups. In Dr group, were...

Abstract A Markovian quantum process can be arbitrarily divided into two or more legitimate completely‐positive (CP) subprocesses. When at least one non‐CP exists among the processes, dynamics is considered non‐Markovian. However, how to utilize minimum experimental efforts, without examining all input states and using entanglement resources, identify measure non‐Markovianity still being determined. Herein, a method proposed quantify processes for identifying measuring burden of state...

The present study collects a large amount of HRI-related research studies and analyzes the trends from 2010 to 2021. Through topic modeling technique, our developed ML model is able retrieve dominant factors. preliminary results reveal five important topics, handover, privacy, robot tutor, skin de deformation, trust. Our show in HRI domain can be divided into two general directions, namely technical human aspects regarding use robotic applications. At this point, we are increasing pool...

A sender can prepare a quantum state for remote receiver using preshared entangled pairs, only the sender's single-qubit measurement, and receiver's simple correction informed by sender. It provides resource-efficient advantages over teleportation information. Here, we propose most efficient approach to detect preparation (RSP) based on benefits powered coherence's static resources of shared pairs dynamic both RSP participants input. requires minimum one additional coherence creation...

This research explores the progression of dementia in aging populations, emphasizing role psychosocial factors, through advanced data analytics. Drawing from Taiwan Longitudinal Study Aging dataset, we examined 13,088 observations 4,869 individuals spanning four survey years. Using machine learning approaches, particularly generalized linear mixed-effects model (GLMM) tree model, pinpointed crucial psychological and social determinants affecting trajectory dementia. Our results highlight...

Understanding and inferencing Heterogeneous Treatment Effects (HTE) Conditional Average (CATE) are vital for developing personalized treatment recommendations. Many state-of-the-art approaches achieve inspiring performance in estimating HTE on benchmark datasets or simulation studies. However, the indirect predicting manner complex model architecture reduce interpretability of these approaches. To mitigate gap between predictive heterogeneity interpretability, we introduce Causal Rule Forest...

Machine learning models are often criticized for their black-box nature, raising concerns about applicability in critical decision-making scenarios. Consequently, there is a growing demand interpretable such contexts. In this study, we introduce Model-based Deep Rule Forests (mobDRF), an representation algorithm designed to extract transparent from data. By leveraging IF-THEN rules with multi-level logic expressions, mobDRF enhances the interpretability of existing without compromising...

Quantum networks typically comprise quantum channels, repeaters, and end nodes. Remote state preparation (RSP) allows one node to prepare the states of other nodes remotely. While discord has recently been recognized as necessary for RSP, it does not guarantee practical implementation RSP in surpasses any classical method. Herein, we theoretically introduce experimentally study a resource that call capability. This validates all static dynamic elements required enable genuine where RSP's can...

In this study, we address the interpretability issue in complex, black-box Machine Learning models applied to sequence data. We introduce Model-Based tree Hidden Semi-Markov Model (MOB-HSMM), an inherently interpretable model aimed at detecting high mortality risk events and discovering hidden patterns associated with Intensive Care Units (ICU). This leverages knowledge distilled from Deep Neural Networks (DNN) enhance predictive performance while offering clear explanations. Our...

As information systems continuously produce high volumes of user event log data, efficient detection anomalous activities indicative insider threats becomes crucial. Typical supervised Machine Learning (ML) methods are often labor-intensive and suffer from the constraints costly labeled data with unknown anomaly dependencies. Here we introduce a knowledge distillation ML framework, using multiple binary classifiers as teacher models multi-label model student. Leveraging soft targets models,...

In this study, we address the interpretability issue in complex, black-box Machine Learning models applied to sequence data. We introduce Model-Based tree Hidden Semi-Markov Model (MOB-HSMM), an inherently interpretable model aimed at detecting high mortality risk events and discovering hidden patterns associated with Intensive Care Units (ICU). This leverages knowledge distilled from Deep Neural Networks (DNN) enhance predictive performance while offering clear explanations. Our...

Photon-mediated quantum networks generally consist of channels, repeaters, and end nodes. Remote state preparation (RSP) enables one the nodes to prepare states other remotely. RSP also serves as a deterministic single-photon source for networking communications. Herein, we theoretically experimentally investigate how surpasses any classical emulation without entanglement qubit unitaries. We introduce new type resource, which refer capability, validate all static dynamic elements required...

A Markovian quantum process can be arbitrarily divided into two or more legitimate completely-positive (CP) subprocesses. When at least one non-CP exists among the processes, dynamics is considered non-Markovian. However, how to utilize minimum experimental efforts, without examining all input states and using entanglement resources, identify measure non-Markovianity still being determined. Herein, we propose a method quantify processes for identifying measuring burden of state optimization...