Wenbo Wang

Minesweeper is both an intellectual challenge for human players and an effective testbed for intelligent algorithms. For the standard Minesweeper task on a 9 × 9 grid with 10 mines, a convolutional neural network (CNN)-based supervised learning approach, together with optimized network architecture, state encoding, and dynamic learning rate scheduling, achieves a test win rate of 85.16% (SD = 0.36 percentage points), exceeding the baseline Single-Point (SP) algorithm by more than 10 percentage points. The combined use of dropout and data augmentation achieves a higher test win rate than either technique alone under the evaluated settings. Building on this model, we propose an easy-to-hard staged curriculum learning (SCL) strategy to address the increasing task difficulty caused by higher mine density and the difficulty of training models from scratch. By progressively increasing the number of mines and transferring model parameters from the preceding training stage, the model can inherit previously acquired knowledge and gradually adapt to more complex tasks. Representative SCL models achieved test win rates of 55.85%, 20.38%, and 4.00% on the 14-, 18-, and 22-mine tasks, respectively. In the three-seed ablation study on the 18-mine task using validation-based checkpoint selection, SCL achieved the highest mean test win rate of 21.65%, compared with 20.85% for mixed-density adaptation, 17.08% for direct transfer, and 15.59% for training from scratch. SCL achieved numerically higher test win rates than mixed-density adaptation across all three seeds, although the difference was not statistically significant. However, a constraint satisfaction problem (CSP) solver achieves a higher win rate of 47.17% on the same task, highlighting the advantage of explicit constraint-based reasoning in absolute solving performance. Further analysis of clue-value distributions shows that the frequencies of local patterns change with mine density, providing a structural motivation for the easy-to-hard curriculum. These findings indicate that staged curriculum learning can facilitate knowledge transfer across increasingly difficult Minesweeper tasks, although reliable solving of high-density configurations remains challenging.

Improving sample efficiency and achieving balanced multi-task performance across different tasks remain important research challenges in reinforcement learning. Yie Ar Kung-Fu is formulated as a multi-task reinforcement learning environment. A task-specific neural network architecture, simplified state and action encoding schemes, a reward mechanism, and a dynamic learning-rate schedule are designed for the game environment. Based on Proximal Policy Optimization (PPO), a parallel training method is further proposed. The method aggregates episode samples generated from different opponent environments into a unified sample queue in the order of episode completion and dynamically adjusts the weights of samples from different opponents in the loss function based on their win rates. Experiments were conducted on the AutoDL platform using an NVIDIA RTX 4090 GPU. The proposed method enables joint training across multiple opponent environments. After approximately 15,000 training episodes, the trained agent achieved win rates exceeding 90% against all opponents and attained a level-clear rate of 79%, demonstrating strong game-playing performance and balanced performance across multiple opponent tasks. Furthermore, interdisciplinary analogies are used to interpret how parallel training improves sample efficiency and balanced performance across tasks. The proposed method is simple to implement, stable during training, and readily applicable to other complex multi-task environments.

This paper introduces and proves a novel proposition on the structural closure property of Linear Time-Invariant (LTI) invertible systems. Specifically, it demonstrates that systems composed by arbitrarily combining LTI invertible systems through cascade, parallel, and feedback configurations remain LTI invertible systems. Furthermore, a constructive approach is proposed to reinterpret systems governed by linear constant-coefficient differential equations, proving that under the assumption of causality, such systems are invertible. The theoretical contribution of this work enriches the understanding of invertibility in signal and system theory, while its practical value lies in providing structural support for the design and analysis of complex systems.

The Minesweeper game is modeled as a sequential decision-making task, for which a neural network architecture, state encoding, and reward function were herein designed. Both a Deep Q-Network (DQN) and supervised learning methods were successfully applied to optimize the training of the game. The experiments were conducted on the AutoDL platform using an NVIDIA RTX 3090 GPU for efficient computation. The results showed that in a 6 × 6 grid with four mines, the DQN model achieved an average win rate of 93.3% (standard deviation: 0.77%), while the supervised learning method achieved 91.2% (standard deviation: 0.9%), both outperforming human players and baseline algorithms and demonstrating high intelligence. The mechanisms of the two methods in the Minesweeper task were analyzed, with the reasons for the faster training speed and more stable performance of supervised learning explained from the perspectives of means–ends analysis and feedback control. Although there is room for improvement in sample efficiency and training stability in the DQN model, its greater generalization ability makes it highly promising for application in more complex decision-making tasks.

Double emulsions (W/O/W) with compartmentalized structures have attracted a lot of research interests due to their diverse applications in the food industry. Herein, oil phase of double emulsions was gelled with beeswax (BW), and the effects of BW mass ratios (0–8.0%) on the stability, oral sensation, and flavor release profile of the emulsions were investigated. Rheological tests revealed that the mechanical properties of double emulsions were dependent on the mass ratio of BW. With the increase in BW content, double emulsions showed a higher resistance against deformation, and lower friction coefficient with a smoother mouthfeel. Turbiscan analysis showed that the addition of BW improved the stability of double emulsions during a 14 days’ storage, under freeze–thawed, and osmotic pressure conditions, but it did not improve the heating stability of double emulsions. The addition of BW contributed to lower air-emulsion partition coefficients of flavor (2,3-diacetyl) compared to those without the addition of BW at 20 °C and 37 °C, respectively. Furthermore, the addition of BW and its mass ratio significantly altered the flavor release behavior during the open-bottle storage of double emulsions. The response value of 0% BW dropped sharply on the first day of opening storage, showing a burst release behavior. While a slow and sustained release behavior was observed in double emulsions with 8.0% BW. In conclusion, gelation of the intermediate oil phase of double emulsions significantly enhanced the stability of double emulsions with tunable oral sensation and flavor release by varying the mass ratio of beeswax.

Due to the presence of oils and fats, emulsions undergo lipid oxidation during storage, resulting in deterioration in sensory and nutritional qualities of emulsions. Phenolic antioxidants are widely used in inhibiting lipid oxidation. We found that phenolic antioxidants could form composite nanoparticles with zein and could be utilized to prepare Pickering emulsions with strong oxidative stability. Phenolic antioxidants with different hydrophobicity (octyl gallate, Propyl gallate and gallic acid) were fabricated into nanoparticles with zein (namely, Z/OG, Z/PG, Z/GA), and the oxidative stability of the as-prepared emulsions were much stronger than Tween 20 stabilized emulsions. And the Z/OG stabilized emulsions showed stronger oxidative stability than those stabilized by Z/PG and Z/GA. We also found that the antioxidative efficiency of octyl gallate was significantly higher than propyl gallate and gallic acid in Tween 20 stabilized emulsions. Compared with the Tween 20 stabilized emulsions, the difference in antioxidation efficiency between octyl gallate, propyl gallate and gallic acid was more obvious in the Pickering emulsions due to the combined effect of physical barrier and phenolic antioxidants. The results of this study help us to improve the antioxidant efficiency of phenolic antioxidants via the rational selection of antioxidant and a well-designed Pickering stabilization strategy.

Lipid oxidation is still one of the major food-safety issues associated with the emulsion-based food systems. Engineering the interfacial region is an effective way to improve the oxidative stability of emulsion. Herein, a novel Pickering emulsion with strong oxidative stability was prepared by using zein nanoparticles and Tween 20 as stabilizers (ZPE). The modulation effects of the particle size on the distribution of gallic acid (GA) and the oxidative stability of ZPE were investigated. In the absence of GA, Pickering emulsions stabilized with different sizes of zein nanoparticles showed similar oxidative stability, and the physical barrier effect took the dominant role in retarding lipid oxidation. Moreover, in the presence of GA, ZPE stabilized by zein nanoparticles with the averaged particle size of 130 nm performed stronger oxidation than those stabilized by zein nanoparticles of 70 and 220 nm. Our study revealed that the interfacial concentration of GA (GA_I) was tuned by zein nanoparticles due to the interaction between them, but the difference in the binding affinity between GA and zein nanoparticles was not the dominant factor regulating the (GA_I). It was the interfacial content of zein nanoparticles (Γ), which was affected by the particle size, modulated the (GA_I) and further dominated the oxidative stability of ZPEs. The present study suggested that the potential of thickening the interfacial layer to prevent lipid oxidation was limited, increasing the interfacial concentration of antioxidant by interfacial engineering offered a more efficient alternative.