EF1-Constrained Nash Social Welfare with Identical Additive Valuations: Complexity, Guarantees, and Experiments
研究了在相同加性估值下不可分割物品的分配问题，通过提出PriorityNet深度强化学习框架来保证EF1并最大化Nash社会福利。
0 citationsRead paper
Selected work
Representative Papers
Resolving Envy by Adding Goods with Bounded Supply: A Type-Count Dichotomy and Two-Agent Hardness
This study investigates the computational complexity of achieving envy-free allocations by adding items when item supplies are limited and agents face no individual budget constraints. Focusing on binary additive valuations, the work establishes a sharp dichotomy based on the number of added item types: a polynomial-time algorithm exists for a single item type, whereas the problem becomes NP-complete with just two item types—even weakly NP-complete for only two agents. The approach models envy relations via a system of difference constraints and employs the Bellman-Ford algorithm to compute a minimal feasible extension. Complexity reductions further delineate the hardness across multiple scenarios. This work fully resolves the open two-type case posed by Bentert et al. and establishes the first precise computational complexity boundary for this setting.
0 citationsRead paper
Testing Full Quartet Consistency: Adaptive Reconstruction, Random Verification, and Constant-Query Testability
This work addresses the problem of efficiently determining whether a given complete quartet topology system is induced by some phylogenetic tree or is ε-far from all tree-induced systems. We present the first explicit polynomial-time adaptive one-sided error property tester for this task. Our approach reconstructs a candidate tree via anchored quartet queries and verifies its consistency using uniformly random queries. The adaptive version requires only O(n log n + ε⁻¹ log(1/δ)) queries, while the non-adaptive variant uses C(n−1,3) + O(ε⁻¹ log(1/δ)) queries—both substantially fewer than the input size Θ(n⁴). Notably, our query complexity is theoretically optimal, matching known upper and lower bounds.
0 citationsRead paper
EMBRACE: A Multi-task Framework for Comprehensive Quality Assessment in Cleavage-stage Embryo
This study addresses key challenges in cleavage-stage embryo assessment during in vitro fertilization—namely, inaccurate fragmentation detection, inconsistent developmental stage classification, and subjective symmetry scoring—by proposing the first multitask deep learning framework that integrates interpretable spatial fragment localization with holistic morphological analysis. Built upon a shared ResNet-50 backbone, the model incorporates a multiscale feature fusion module, a U-Net–style segmentation decoder, and task-specific classification heads to simultaneously perform cytoplasmic fragment segmentation, t2/t4 stage classification, and blastomere symmetry grading. On an independent test set, the framework achieves a Dice coefficient of 0.781 for segmentation, 0.995 accuracy for developmental stage classification, and 0.901 balanced accuracy for symmetry grading (weighted Kappa = 0.859), substantially enhancing the objectivity and consistency of embryo evaluation.
0 citationsRead paper
RL-Ballast: Ship Ballast Water Path Planning and Clog Prediction via Reinforcement Learning
This study addresses the limited adaptability of conventional ship ballast water systems under hydraulic anomalies—such as valve failures or pipe blockages—and their heavy reliance on dense sensor arrays for fault diagnosis. The authors propose a novel approach that integrates graph theory with deep reinforcement learning, modeling ballast routing as a set of 54 feasible fluid transfer paths. By employing frame-stacked water level observations and action outcomes to approximate a partially observable environment, the method incorporates failure-action memory and dynamic action masking to enable adaptive rerouting. Notably, it implicitly infers blockage states without explicit high-dimensional POMDP modeling and introduces a fault-history scoring mechanism reliant only on sparse sensing to rank suspect components. Experimental results demonstrate 100% task success across all single-point blockage scenarios, reducing average decision steps from 61.0 to 41.5; the fault-scoring mechanism achieves 100% top-3 hit rate, with strict and inclusive top-1 hit rates of 66.7% and 83.3%, respectively.
0 citationsRead paper
Recent publications
Latest Papers
EF1-Constrained Nash Social Welfare with Identical Additive Valuations: Complexity, Guarantees, and Experiments
研究了在相同加性估值下不可分割物品的分配问题，通过提出PriorityNet深度强化学习框架来保证EF1并最大化Nash社会福利。
0 citationsRead paper
Resolving Envy by Adding Goods with Bounded Supply: A Type-Count Dichotomy and Two-Agent Hardness
This study investigates the computational complexity of achieving envy-free allocations by adding items when item supplies are limited and agents face no individual budget constraints. Focusing on binary additive valuations, the work establishes a sharp dichotomy based on the number of added item types: a polynomial-time algorithm exists for a single item type, whereas the problem becomes NP-complete with just two item types—even weakly NP-complete for only two agents. The approach models envy relations via a system of difference constraints and employs the Bellman-Ford algorithm to compute a minimal feasible extension. Complexity reductions further delineate the hardness across multiple scenarios. This work fully resolves the open two-type case posed by Bentert et al. and establishes the first precise computational complexity boundary for this setting.
0 citationsRead paper
Testing Full Quartet Consistency: Adaptive Reconstruction, Random Verification, and Constant-Query Testability
This work addresses the problem of efficiently determining whether a given complete quartet topology system is induced by some phylogenetic tree or is ε-far from all tree-induced systems. We present the first explicit polynomial-time adaptive one-sided error property tester for this task. Our approach reconstructs a candidate tree via anchored quartet queries and verifies its consistency using uniformly random queries. The adaptive version requires only O(n log n + ε⁻¹ log(1/δ)) queries, while the non-adaptive variant uses C(n−1,3) + O(ε⁻¹ log(1/δ)) queries—both substantially fewer than the input size Θ(n⁴). Notably, our query complexity is theoretically optimal, matching known upper and lower bounds.
0 citationsRead paper
EMBRACE: A Multi-task Framework for Comprehensive Quality Assessment in Cleavage-stage Embryo
This study addresses key challenges in cleavage-stage embryo assessment during in vitro fertilization—namely, inaccurate fragmentation detection, inconsistent developmental stage classification, and subjective symmetry scoring—by proposing the first multitask deep learning framework that integrates interpretable spatial fragment localization with holistic morphological analysis. Built upon a shared ResNet-50 backbone, the model incorporates a multiscale feature fusion module, a U-Net–style segmentation decoder, and task-specific classification heads to simultaneously perform cytoplasmic fragment segmentation, t2/t4 stage classification, and blastomere symmetry grading. On an independent test set, the framework achieves a Dice coefficient of 0.781 for segmentation, 0.995 accuracy for developmental stage classification, and 0.901 balanced accuracy for symmetry grading (weighted Kappa = 0.859), substantially enhancing the objectivity and consistency of embryo evaluation.
0 citationsRead paper
RL-Ballast: Ship Ballast Water Path Planning and Clog Prediction via Reinforcement Learning
This study addresses the limited adaptability of conventional ship ballast water systems under hydraulic anomalies—such as valve failures or pipe blockages—and their heavy reliance on dense sensor arrays for fault diagnosis. The authors propose a novel approach that integrates graph theory with deep reinforcement learning, modeling ballast routing as a set of 54 feasible fluid transfer paths. By employing frame-stacked water level observations and action outcomes to approximate a partially observable environment, the method incorporates failure-action memory and dynamic action masking to enable adaptive rerouting. Notably, it implicitly infers blockage states without explicit high-dimensional POMDP modeling and introduces a fault-history scoring mechanism reliant only on sparse sensing to rank suspect components. Experimental results demonstrate 100% task success across all single-point blockage scenarios, reducing average decision steps from 61.0 to 41.5; the fault-scoring mechanism achieves 100% top-3 hit rate, with strict and inclusive top-1 hit rates of 66.7% and 83.3%, respectively.
0 citationsRead paper