DeepLogic: Joint Learning of Neural Perception and Logical Reasoning.

Neural-symbolic learning, aiming to combine the perceiving power of neural perception and the reasoning power of symbolic logic together, has drawn increasing research attention. However, existing works simply cascade the two components together and optimize them isolatedly, failing to utilize the mutual enhancing information between them. To address this problem, we propose DeepLogic, a framework with joint learning of neural perception and logical reasoning, such that these two components are jointly optimized through mutual supervision signals. In particular, the proposed DeepLogic framework contains a deep-logic module that is capable of representing complex first-order-logic formulas in a tree structure with basic logic operators. We then theoretically quantify the mutual supervision signals and propose the deep&logic optimization algorithm for joint optimization. We further prove the convergence of DeepLogic and conduct extensive experiments on model performance, convergence, and generalization, as well as its extension to the continuous domain. The experimental results show that through jointly learning both perceptual ability and logic formulas in a weakly supervised manner, our proposed DeepLogic framework can significantly outperform DNN-based baselines by a great margin and beat other strong baselines without out-of-box tools.

Structure and activity alteration in adult highland residents' cerebrum: Voxel-based morphometry and amplitude of low-frequency fluctuation study.

Introduction: People living in highland areas may have factors that allow them to adapt to chronic hypoxia, but these physiological mechanisms remain unclear. This study aimed to investigate the brain mechanism in a cohort of adult residents of Tibet, a well-known plateau section in China, by observing differences in brain structure and function in non-plateau populations. Methods: The study included 27 Tibetan and 27 non-plateau region residents who were matched in age, sex, and education. All participants underwent high-resolution three-dimensional T1 weighted imaging (3D-T1WI) and resting-state functional magnetic resonance imaging (rs-fMRI) scans on a 1.5 Tesla MR. Gray matter volumes and regional spontaneous neuronal activity (SNA) were calculated and compared between the two groups. Results: When comparing gray matter in people living in high altitudes to those living in the flatlands, the results showed positive activation of gray matter in local brain regions (p < 0.05, false discovery rate (FDR) corrected), in the right postcentral [automated atomic labeling (aal)], left postcentral (aal), and right lingual (aal) regions. Comparing the people of high altitude vs. flat land in the brain function study (p < 0.05, FDR corrected), positive activation was found in the right superior motor area (aal) and left superior frontal (aal), and negative activation was found in the right precuneus (aal). Conclusion: In high-altitude individuals, larger regional gray matter volumes and higher SNA may represent a compensatory mechanism to adapt to chronic hypoxia.