Digital Commons Network^™

Automated Assignment Grading With Large Language Models: Insights From A Bioinformatics Course, Pavlin G Poličar, Martin Špendl, Tomaž Curk, Blaž Zupan

Faculty, Staff and Students Publications

Motivation: Providing students with individualized feedback through assignments is a cornerstone of education that supports their learning and development. Studies have shown that timely, high-quality feedback plays a critical role in improving learning outcomes. However, providing personalized feedback on a large scale in classes with large numbers of students is often impractical due to the significant time and effort required. Recent advances in natural language processing and large language models (LLMs) offer a promising solution by enabling the efficient delivery of personalized feedback. These technologies can reduce the workload of course staff while improving student satisfaction and learning outcomes. Their …

Cardiac Rhythm Devices In Heart Failure With Reduced Ejection Fraction – Role, Timing, And Optimal Use In Contemporary Practice European Journal Of Heart Failure Expert Consensus Document, Biykem Bozkurt, Wilfried Mullens, Christophe Leclercq, Andrea M Russo, Gianluigi Savarese, Michael Böhm, Loreena Hill, Koichiro Kinugawa, Naoki Sato, William T Abraham, Antoni Bayes-Genis, Alexandre Mebazaa, Giuseppe M C Rosano, Shelley Zieroth, Cecilia Linde, Javed Butler

Faculty, Staff and Students Publications

Guidelines for management of heart failure with reduced ejection fraction (HFrEF) emphasize personalized care, patient engagement, and shared decision-making. Medications and cardiac rhythm management (CRM) devices are recommended with a high level of evidence. However, there are significant disparities: patients who could benefit from devices are frequently referred too late or not at all. Misconceptions about device therapy and the notion that the needs of patients (especially the prevention of sudden cardiac death) can now be met by expanding drug therapies may play a role in these disparities. This state-of-the-art review is produced by members of the DIRECT HF initiative, …

Development And Diversity Of Astrocytes, Michael R Williamson, Sanjana Murali, Benjamin Deneen

Faculty, Staff and Students Publications

Proper development of the central nervous system (CNS) depends on astrocytes, which are the most abundant glial cells in the CNS. Here, we provide an overview of astrocyte ontogeny and morphogenesis, highlighting mechanisms that regulate astrocyte development. In addition, we describe the roles of astrocytes in regulating the development of other CNS cell types, with emphasis on neural circuits and vascular networks. We also discuss evidence for astrocyte diversity and its implications for CNS development. Finally, we pose open questions related to astrocyte development and its intersection with astrocyte diversity.

Clinical Assessment On Days 1-14 For The Characterization Of Traumatic Brain Injury: Recommendations From The 2024 Ninds Traumatic Brain Injury Classification And Nomenclature Initiative Clinical/Symptoms Working Group, David K Menon, Noah D Silverberg, Adam R Ferguson, Thomas J Bayuk, Shubhayu Bhattacharyay, David L Brody, Scott A Cota, Ari Ercole, Anthony Figaji, Guoyi Gao, Christopher C Giza, Fiona Lecky, Rebekah Mannix, Ana Mikolić, Kasey E Moritz, Claudia S Robertson, Abel Torres-Espin, Spyridoula Tsetsou, John K Yue, Hibah O Awad, Kristen Dams-O'Connor, Adele Doperalski, Andrew I R Maas, Michael A Mccrea, Nsini Umoh, Geoffrey T Manley

Faculty, Staff and Students Publications

The current classification of traumatic brain injury (TBI) primarily uses the Glasgow Coma Scale (GCS) to categorize injuries as mild (GCS 13-15), moderate (GCS 9-12), or severe (GCS ≤8). However, this system is unsatisfactory, as it overlooks variations in injury severity, clinical needs, and prognosis. A recent report by the National Academies of Sciences, Engineering, and Medicine (USA) recommended updating the classification system, leading to a workshop in 2024 by the National Institute of Neurological Disorders and Stroke. This resulted in the development of a new clinical, biomarker, imaging, and modifier (CBI-M) framework, with input from six working groups, including …

Predictors Of Failure To Rescue After Fenestrated-Branched Endovascular Aortic Repair Of Thoracoabdominal Aortic Aneurysms: Presented At The Thirty-Seventh Annual Meeting Of The European Society Of Vascular Surgery, Belfast, Northern Ireland, September 26-29, 2023, Andrea Vacirca, Thomas Mesnard, Ying Huang, Bernardo C Mendes, Tomasz Jakimowicz, Darren B Schneider, Stéphan Haulon, Jonathan Sobocinski, Adam W Beck, Andres Schanzer, Mark A Farber, Carlos Timaran, Andrea Kahlberg, Tilo Kölbel, Warren J Gasper, Barend M E Mees, Mauro Gargiulo, Nuno V Dias, Anthony Lee Woongchae, Matthew P Sweet, Kevin Mani, Matthew Eagleton, Luis Mendes Pedro, Hence Verhagen, Kak Khee Yeung, Nikolaos Tsilimparis, Timothy Resch, Luca Bertoglio, Emília Ferreira, Manar Khashram, Titia Sulzer, Marina Dias-Neto, Emanuel R Tenorio, Lucas Ruiter Kanamori, Katarzyna Jama, Ezequiel Parodi, Vivian Gomes, Jesus Porras Colon, Roberto Chiesa, Giuseppe Panuccio, Geert Willem Schurink, Charlotte Lemmens, Enrico Gallitto, Gianluca Faggioli, Angelos Karelis, Anders Wanhainen, Mohammed Habib, Ryan Gouveia E Melo, Kaj Olav Kappe, Samira Elize Mariko Van Knippenberg, Bichlan Tran, Sean Crawford, Marco Panagrosso, Andrea Melloni, Stefano Bonardelli, Rita Garcia, Tiago Ribeiro, Sinead Gormley, Steven Maximus, Gustavo S Oderich, International Multicenter Aortic Research Group

Faculty, Staff and Students Publications

No abstract provided.

Role Of Kctd13 In Modulating Ar And Sox9 Expression In Different Penile Cell Populations, Carolina J Jorgez, Ahmed Chahdi, Hunter Flores, Marisol O'Neill, Abhishek Seth

Faculty, Staff and Students Publications

Objective: Micropenis is a condition with significant physical and psychological implications caused mainly by decreased androgen action in penile development. Kctd13-knockout (Kctd13-KO) mice have micropenis, cryptorchidism, and fertility defects because of reduced levels of androgen receptor (AR) and SOX9. We hypothesized that normalizing the levels of AR and SOX9 in the Kctd13-KO penis could help us to understand the mechanism of action of these signaling pathways on penile development.

Methods: We generated transgenic mice lacking Kctd13 and conditionally expressing AR in the urethral mesenchyme after Cre activation with Twist2cre (Kctd13-KO; AR-CMV; Twist2cre; herein called AR+), and Sox9 in the urethral …

Immediate Skin-To-Skin Contact And Postpartum Hemorrhagic Morbidity, Rachel L Wiley, Ipsita Ghose, Dana R Canfield, Minhazur R Sarker, Hector Mendez-Figueroa, Suneet Chauhan

Faculty, Staff and Students Publications

Objective: To examine rates of postpartum hemorrhagic (PPH) morbidity among patients who did and did not have immediate skin-to-skin contact (SSC).

Methods: This study was a retrospective cohort of all non-anomalous, term singleton vaginal births at a Level IV center over 2 years. Exclusion criteria included COVID-19. Immediate SSC was defined as at least 60 min of direct contact initiated between parturient and neonate within 10 min of birth. The primary outcome was a composite of maternal morbidity related to PPH compared among those with and without immediate SSC. We used multivariable Poisson regression adjusted for possible confounders with robust …

Formation Of An Expanding Memory Representation In The Hippocampus, Sachin P Vaidya, Guanchun Li, Raymond A Chitwood, Yiding Li, Jeffrey C Magee

Faculty, Staff and Students Publications

How brain networks connected by labile synapses store new information without catastrophically overwriting previous memories remains poorly understood. To examine this, we tracked the same population of hippocampal CA1 place cells (PCs) as mice learned a task for 7 days. We found evidence of memory formation as both the number of PCs maintaining a stable place field and the stability of individual PCs progressively increased across the week until most of the representation was composed of long-term stable PCs. The stable PCs disproportionately represented task-related learned information, were retrieved earlier within a behavioral session and showed a strong correlation with …

Undominated Mechanisms, Tilman Borgers, Jiangtao Li, Kexin Wang

Research Collection School Of Economics

We study the design of mechanisms when the designer faces multiple plausible scenarios and is uncertain about the true scenario. A mechanism is dominated by another if the latter performs at least as well in all plausible scenarios and strictly better in at least one. A mechanism is undominated if no other feasible mechanism dominates it. We show how analyzing undominated mechanisms could be useful and illustrate the tractability of characterizing such mechanisms. This approach provides an alternative criterion for mechanism design under non-Bayesian uncertainty, complementing existing methods.

Do Search Costs Explain Persistent Investment In Active Mutual Funds?, Aljoscha Janssen, Jurre Thiel

Research Collection School Of Economics

Active funds, though losing market share since the 1990s, make up nearly half of all mutual funds but charge more without better performance. We analyze fund data and a search model, highlighting the impact of search costs and active fund preferences. From 1993 to 2018, reduced search costs expanded the market and heightened competition, while a preference shift from active to passive funds increased the latter's market share. However, investors who choose active funds, facing higher search costs, and continue to show a strong preference for them, allow these funds to keep charging higher fees.

Technifying Ventures, Yoshiki Ando, Emin Dinlersoz, Jeremy Greenwood, Ruben Piazzesi

Research Collection School Of Economics

How do advanced technology adoption and venture capital (VC) funding impact employment and growth? An analysis of data from the US Census Bureau suggests that while both advanced technology use and VC funding matter on their own for firm outcomes, their joint presence is most strongly correlated with higher employment levels. VC presence is linked with a high increase in employment, though primarily among a limited subset of firms. In contrast, technology adoption is associated with a smaller rise in employment, yet it influences a considerably larger number of firms. A model of startups is created, focusing on decisions to …

The Rising Returns To R&D: Ideas Are Not Getting Harder To Find, Yoshiki Ando, James Bessen, Xiupeng Wang

Research Collection School Of Economics

R&D has grown robustly, yet aggregate productivity growth has stagnated. Is this because “ideas are getting harder to find”? Using microdata from the US Census Bureau this paper estimates R&D effectiveness in the manufacturing sector from 1976 to 2018. We find that the marginal effect of R&D on output and the gross marginal returns to R&D have risen sharply. But rivals’ R&D has also gotten more effective, raising expected obsolescence. But firm productivity growth rises with growth in the R&D stock. Higher obsolescence slows the accumulation of R&D capital, reducing the contribution to productivity growth and shortening innovation lives.

Public Spaces For Community Campuses And Universities, Marichela Sepe

Books and Monographs

Chapter: Modern Pedestrian Campus Design in North America, 1940s-70s, Eric Mumford, pages 13-22

This volume explores the growing significance of open public spaces within università campuses—spaces that go far beyond their functional role to become essential arenas for social interaction, informal learning, well-being, and urban integration.

Structured around three interconnected themes—Campus and Public Spaces, Campus and University Communities, and Campus and the City—the volume gathers international best practices, offering an insightful tool for students, academics, designers, administrators, and urban thinkers.

Full text also available for download at https://www.inuedizioni.com/en/prodotti/pubblicazione/public-spaces-community-campuses-and-universities

Digital Economy Innovation In The Indo-Pacific: Towards A 'Singapore Effect'?, J.G. Allen, Qiu Xu Martin Liao

Research Collection Yong Pung How School Of Law

This paper examines the rise of Digital Economy Agreements (DEAs) in the Indo-Pacific as a strategic response to digital trade fragmentation and great power competition. Focusing on Singapore’s leadership, we introduce the ‘Singapore Effect’ – a model of regulatory influence distinct from others, such as the ‘Brussels Effect’ and ‘Beijing Effect.’ Unlike market-driven regulatory diffusion, the Singapore Effect emphasizes interoperability, norm entrepreneurship, and mini-lateralism, enabling middle powers to shape digital trade governance. We analyze DEA formation, implementation challenges in national law, and Singapore’s role as a bridge between diverse regulatory regimes. DEAs’ flexible, modular structure allows for incremental regulatory alignment …

The B2biers System: A Content-Based Perspective On Maximizing Influence And Subscription In Social Networks, Konstantinos Theocharidis, Hady Wirawan Lauw

Research Collection School Of Computing and Information Systems

The popular problem of Influence Maximization (IM) asks for the k users who can maximize the influence of a fixed post in a social network. In contrast, the problem of Content- Aware Influence Maximization (CAIM) asks for the k features to form a viral tunable post in a social network starting its diffusion from a fixed set of initial adopters. CAIM paves the way for a number of novel problems to be studied that altogether can lead to the development of a system that would be valuable for advertisers who manage social network pages. This holds since features (brands) in …

Evaluating Chatgpt To Answer Multi-Modal Exercises In Computer Science Education, Eng Lieh Ouh, Kar Way Tan, Siaw Ling Lo, Benjamin Gan

Research Collection School Of Computing and Information Systems

This study investigates ChatGPT-4o's ability to answer multi-modal assessment exercises in computer science (CS) courses. While the use of large language models (LLMs) to answer text-based exercises are extensively researched, their ability to answer exercises involving artifacts of other modalities remains underexplored. To close this gap, we evaluate ChatGPT-4o's answers to 120 multi-modal CS exercises in programming, software design, human-computer interaction, statistical analysis, process analysis, and simulation. The multi-modal artifacts in these exercises include class diagrams, sequence diagrams, user interface images, analytical charts, workflow diagrams and object-flow diagrams. Our comparisons to the expected answers of these exercises show that ChatGPT-4o …

Position: Trustworthy Ai Agents Require The Integration Of Large Language Models And Formal Methods, Yedi Zhang, Yufan Cai, Xinyue Zuo, Xiaokun Luan, Kailong Wang, Zhe Hou, Yifan Zhang, Zhiyuan Wei, Meng Sun, Jun Sun, Jing Sun, Jin Song Dong

Research Collection School Of Computing and Information Systems

Large Language Models (LLMs) have emerged as a transformative AI paradigm, profoundly influencing broad aspects of daily life. Despite their remarkable performance, LLMs exhibit a fundamental limitation: hallucination—the tendency to produce misleading outputs that appear plausible. This inherent unreliability poses significant risks, particularly in high-stakes domains where trustworthiness is essential. On the other hand, Formal Methods (FMs), which share foundations with symbolic AI, provide mathematically rigorous techniques for modeling, specifying, reasoning, and verifying the correctness of systems. These methods have been widely employed in mission-critical domains such as aerospace, defense, and cybersecurity. However, the broader adoption of FMs remains constrained …

An On-The-Fly Synthesis Framework For Ltl Over Finite Traces, Shengping Xiao, Yongkang Li, Shufang Zhu, Jun Sun, Jianwen Li, Geguang Pu, Moshe Vardi

Research Collection School Of Computing and Information Systems

We present an on-the-fly synthesis framework for Linear Temporal Logic over finite traces (LTLf) based on top-down deterministic automata construction. Existing approaches rely on constructing a complete Deterministic Finite Automaton (DFA) corresponding to the LTLf specification, a process with doubly exponential complexity relative to formula size in the worst case. In this case, the synthesis cannot be conducted until the entire DFA is constructed. This inefficiency is the main bottleneck of existing approaches. To address this challenge, we first present a method for converting LTLf into Transition-based DFA (TDFA) by directly leveraging LTLf semantics, incorporating intermediate results as direct components …

Unified Neural Backdoor Removal With Only Few Clean Samples Through Unlearning And Relearning, Nay Myat Min, Long H. Pham, Jun Sun

Research Collection School Of Computing and Information Systems

Deep neural networks have achieved remarkable success across various applications; however, their vulnerability to backdoor attacks poses severe security risks—especially in situations where only a limited set of clean samples is available for defense. In this work, we address this critical challenge by proposing ULRL (UnLearn and ReLearn for backdoor removal), a novel two-phase approach for comprehensive backdoor removal. Our method first employs an unlearning phase, in which the network’s loss is intentionally maximized on a small clean dataset to expose neurons that are excessively sensitive to backdoor triggers. Subsequently, in the relearning phase, these suspicious neurons are recalibrated using …

Advancing Food Nutrition Estimation Via Visual-Ingredient Feature Fusion, Huiyan Qi, Bin Zhu, Chong-Wah Ngo, Jingjing Chen, Ee-Peng Lim

Research Collection School Of Computing and Information Systems

Nutrition estimation is an important component of promoting healthy eating and mitigating diet-related health risks. Despite advances in tasks such as food classification and ingredient recognition, progress in nutrition estimation is limited due to the lack of datasets with nutritional annotations. To address this issue, we introduce FastFood, a dataset with 84,446 images across 908 fast food categories, featuring ingredient and nutritional annotations. In addition, we propose a new model-agnostic Visual-Ingredient Feature Fusion (VIF2 ) method to enhance nutrition estimation by integrating visual and ingredient features. Ingredient robustness is improved through synonym replacement and resampling strategies during training. The ingredient-aware …

Query Understanding In Llm-Based Conversational Information Seeking, Yifei Yuan, Zahra Abbasiantaeb, Mohammad Aliannejadi, Yang Deng

Research Collection School Of Computing and Information Systems

Query understanding in CIS involves accurately interpreting user intent through context-aware interactions. This includes resolving ambiguities, refining queries, and adapting to evolving information needs. LLM enhance this process by interpreting nuanced language and adapting dynamically, improving the relevance and precision of search results in real-time. In this tutorial, we explore advanced techniques to enhance query understanding in LLM-based CIS systems. We delve into LLM-driven methods for developing robust evaluation metrics to assess query understanding quality in multi-turn interactions, strategies for building more interactive systems, and applications like proactive query management and query reformulation. We also discuss key challenges in integrating …

Sanitizable Cross-Domain Access Control With Policy-Driven Dynamic Authorization, Jianfei Sun, Guowen Xu, Hongwei Li, Tianwei Zhang, Cong Wu, Xuehuan Yang, Robert H. Deng

Research Collection School Of Computing and Information Systems

The increasing demand for secure and efficient data sharing has underscored the importance of developing robust cryptographic schemes. However, many existing endeavors have overlooked the following critical issues: (1) unauthorized access resulting from malicious information leakage by senders; (2) absence of constraints on write and read permissions for participants; (3) and inflexibility of strategies to dynamically designate ciphertexts to multiple recipients. In this paper, we present SCPA, a cross-domain access control scheme imbued with sanitization features and propelled by policy-driven dynamic authorization, tailored for cloud-based data sharing. This scheme not only facilitates access controls, including regulations for no-read and no-write …

Action Dubber: Timing Audible Actions Via Inflectional Flow, Wenlong Wan, Weiying Zheng, Tianyi Xiang, Guiqing Li, Shengfeng He

Research Collection School Of Computing and Information Systems

We introduce the task of Audible Action Temporal Localization, which aims to identify the spatiotemporal coordinates of audible movements. Unlike conventional tasks such as action recognition and temporal action localization, which broadly analyze video content, our task focuses on the distinct kinematic dynamics of audible actions. It is based on the premise that key actions are driven by inflectional movements; for example, collisions that produce sound often involve abrupt changes in motion. To capture this, we propose T A2Net, a novel architecture that estimates inflectional flow using the second derivative of motion to determine collision timings without relying on audio …

Unambiguous Granularity Distillation For Asymmetric Image Retrieval, Hongrui Zhang, Yi Xie, Haoquan Zhang, Cheng Xu, Xuandi Luo, Donglei Chen, Xuemiao Xu, Huaidong Zhang, Pheng Ann Heng, Shengfeng He

Research Collection School Of Computing and Information Systems

Previous asymmetric image retrieval methods based on knowledge distillation have primarily focused on aligning the global features of two networks to transfer global semantic information from the gallery network to the query network. However, these methods often fail to effectively transfer local semantic information, limiting the fine-grained alignment of feature representation spaces between the two networks. To overcome this limitation, we propose a novel approach called Layered-Granularity Localized Distillation (GranDist). GranDist constructs layered feature representations that balance the richness of contextual information with the granularity of local features. As we progress through the layers, the contextual information becomes more detailed, …

Leakage-Resilient Easily Deployable And Efficiently Searchable Encryption (Edese), Jiaming Yuan, Yingjiu Li, Jun Li, Daoyuan Wu, Jianting Ning, Yangguang Tian, Robert H. Deng

Research Collection School Of Computing and Information Systems

Easily Deployable and Efficiently Searchable Encryption (EDESE) is a cryptographic primitive designed for practical searchable applications, offering efficient search and easy deployment. However, it remains vulnerable to Leakage-Abuse attacks, allowing adversaries to exploit keyword-matching processes to extract sensitive information. To address these vulnerabilities, we introduce Leakage-Resilient EDESE (LR-EDESE) with k-indistinguishability and controlled leakage functions. We then propose Volume Leakage-Resilient EDESE (VLR-EDESE), a new scheme to protect against both query and document volume leakage. Our experimental results demonstrate that at k = 5000 (maximum security setting), VLR-EDESE incurs an overhead of 63× compared to the baseline EDESE without leakage protection, outperforming …

Llm-Based Multi-Agent Systems For Software Engineering: Literature Review, Vision And The Road Ahead, Junda He, Christoph Treude, David Lo

Research Collection School Of Computing and Information Systems

Integrating Large Language Models (LLMs) into autonomous agents marks a significant shift in the research landscape by offering cognitive abilities that are competitive with human planning and reasoning. This paper explores the transformative potential of integrating Large Language Models into Multi-Agent (LMA) systems for addressing complex challenges in software engineering (SE). By leveraging the collaborative and specialized abilities of multiple agents, LMA systems enable autonomous problem-solving, improve robustness, and provide scalable solutions for managing the complexity of real-world software projects. In this paper, we conduct a systematic review of recent primary studies to map the current landscape of LMA applications …

Surrogate Learning In Meta-Black-Box Optimization: A Preliminary Study, Zeyuan Ma, Zhiyang Huang, Jiacheng Chen, Zhiguang Cao, Yue-Jiao Gong

Research Collection School Of Computing and Information Systems

Recent Meta-Black-Box Optimization (MetaBBO) approaches have shown possibility of enhancing the optimization performance through learning meta-level policies to dynamically configure low-level optimizers. However, existing MetaBBO approaches potentially consume massive function evaluations to train their meta-level policies. Inspired by the recent trend of using surrogate models for cost-friendly evaluation of expensive optimization problems, in this paper, we propose a novel MetaBBO framework which combines surrogate learning process and reinforcement learning-aided Differential Evolution algorithm, namely Surr-RLDE, to address the intensive function evaluation in MetaBBO. Surr-RLDE comprises two learning stages: surrogate learning and policy learning. In surrogate learning, we train a Kolmogorov-Arnold Networks …

A Mixed-Curvature Based Pre-Training Paradigm For Multi-Task Vehicle Routing Solver, Suyu Liu, Zhiguang Cao, Shanshan Feng, Yew-Soon Ong

Research Collection School Of Computing and Information Systems

Solving various types of vehicle routing problems (VRPs) using a unified neural solver has garnered significant attentions in recent years. Despite their effectiveness, existing neural multi-task solvers often fail to account for the geometric structures inherent in different tasks, which may result in suboptimal performance. To address this limitation, we propose a curvature-aware pre-training framework. Specifically, we leverage mixed-curvature spaces during the feature fusion stage, encouraging the model to capture the underlying geometric properties of each instance. Through extensive experiments, we evaluate the proposed pre-training strategy on existing neural multi-task solvers across a variety of testing scenarios. The results demonstrate …

Meta-Black-Box-Optimization Through Offline Q-Function Learning, Zeyuan Ma, Zhiguang Cao, Zhou Jiang, Hongshu Guo, Yue-Jiao Gong

Research Collection School Of Computing and Information Systems

Recent progress in Meta-Black-Box-Optimization (MetaBBO) has demonstrated that using RL to learn a meta-level policy for dynamic algorithm configuration (DAC) over an optimization task distribution could significantly enhance the performance of the low-level BBO algorithm. However, the online learning paradigms in existing works makes the efficiency of MetaBBO problematic. To address this, we propose an offline learning-based MetaBBO framework in this paper, termed Q-Mamba, to attain both effectiveness and efficiency in MetaBBO. Specifically, we first transform DAC task into long-sequence decision process. This allows us further introduce an effective Q-function decomposition mechanism to reduce the learning difficulty within the intricate …

Shield: Multi-Task Multi-Distribution Vehicle Routing Solver With Sparsity And Hierarchy, Yong Liang Goh, Zhiguang Cao, Yining Ma, Jianan Zhou, Mohammed Haroon Dupty, Wee Sun Lee

Research Collection School Of Computing and Information Systems

Recent advances toward foundation models for routing problems have shown great potential of a unified deep model for various VRP variants. However, they overlook the complex real-world customer distributions. In this work, we advance the Multi-Task VRP (MTVRP) setting to the more realistic yet challenging Multi-Task Multi-Distribution VRP (MTMDVRP) setting, and introduce SHIELD, a novel model that leverages both sparsity and hierarchy principles. Building on a deeper decoder architecture, we first incorporate the Mixture-of-Depths (MoD) technique to enforce sparsity. This improves both efficiency and generalization by allowing the model to dynamically select nodes to use or skip each decoder layer, …