Search

Reinforcement learning (RL) for real-time security constrained economic dispatch (RT-SCED) problems have been the subject of significant research interest in recent years. However, ordinary RL approaches struggle to ensure satisfaction of system- and device-wise constraints, having to penalize constraint violations individually. With increasing penetration of renewable energy sources, large-scale energy storage integration is witnessed, driven by their ability to mitigate RES intermittency. This gives rise to the need of time-coupling constraint satisfaction in the RT-SCED problems. Existing safe RL methods either rectify unsafe actions at each time step with a safety layer, which may lead to sub-optimal actions devised at the boundary of feasible space, and may violate time-coupling constraints; or construct a safety evaluation model, which may violate single-step constraints. To address these limitations, this paper proposes a novel safe deep RL method, featuring safety exploration and safety optimization modules, facilitating comprehensive satisfaction of single-step and time-coupling constraints. Furthermore, the policy network features a residual network architecture and allows direct computation of real-value dispatch of all controllable resources, adapting to their distinct power output ranges. Case studies validate the effectiveness of the proposed method in cost efficiency, operational security, computational and scalability performance, compared to state-of-the-art model-driven and data-driven baseline methods, on the IEEE 39-bus and 118-bus test systems.

Although deep reinforcement learning based on Markov Decision Process (MDP) constitutes a well-suited method for real-time control under uncertainties, its application to stochastic dynamic optimal power flow (SDOPF) problem is still challenging in the presence of increasing proliferation of various distributed energy resources, driven by its limitations on constraints satisfaction under uncertainties. While pioneering research explored Constrained MDP and Risk-Aware MDP formulations of SDOPF pursuing cumulative constraint violation minimization, they both struggle with satisfaction of state-wise safety constraints. This letter proposes a Chance Constrained MDP formulation of SDOPF and a Bayesian advantage policy optimization solution method. Bayesian neural networks are used to construct the probability distributions of state- and trajectory-wise constraint violations, and a novel advantage function is incorporated to improve both the policy's quality and safety. Case studies validated the cost-efficiency and comprehensive safety performance of the proposed method against the state-of-the-art.

The microstructural evolution, precipitation behavior, and hardness variation of the Ti-V-Mo complex microalloyed steel during isothermal holding following austenite deformation have been investigated using thermal simulation tester (Gleeble-3800), optical microscope (OM), scanning electron microscope (SEM), electron backscatter diffraction (EBSD), transmission electron microscope (TEM), and Vickers hardness tester. The results indicate that the microstructure transforms from austenite to ferrite as the holding time increases from 0 s to 5400 s at 630 °C, with the microstructure primarily comprising martensite and ferrite. The proportion of ferrite continues to increase, while the proportion of martensite decreases. At a holding time of 600 s, the austenite has nearly completed the ferrite transformation (95%), and the average grain size is very fine (3.9 μm). At this point, the (Ti, V, Mo)C particles exhibit the highest density and a smaller size (7.7 nm), with the hardness reaching its maximum value (415 HV). Concurrently, numerous spherical precipitates particles with a size of < 10 nm formed in the ferrite region, conforming to the Baker–Nutting (BN) orientation relationship with respect to the matrix: (100)(Ti, V, Mo)C//(100)α-Fe, [011](Ti, V, Mo)C//[012]α-Fe. As the holding time increases from 30 s to 60 s, the hardness decreases rapidly due to the reduction of martensite and addition of soft-phase ferrite. However, the subsequent slow decrease in hardness is primarily attributed to weakened precipitation strengthening caused by precipitates coarsening and decreased grain refinement strengthening due to the slight increase in ferrite grain size as the holding time increases from 600 s to 5400 s.

The probability prediction of multivariate time series is a notoriously challenging but practical task. On the one hand, the challenge is how to effectively capture the cross-series correlations between interacting time series, to achieve accurate distribution modeling. On the other hand, we should consider how to capture the contextual information within time series more accurately to model multivariate temporal dynamics of time series. In this work, we proposed a novel non-autoregressive deep learning model, called Multi-scale Attention Normalizing Flow(MANF), where we combine multi-scale attention with relative position information and the multivariate data distribution is represented by the conditioned normalizing flow. Additionally, compared with autoregressive modeling methods, our model avoids the influence of cumulative error and does not increase the time complexity. Extensive experiments demonstrate that our model achieves state-of-the-art performance on many popular multivariate datasets.

Binding affinity prediction of three-dimensional (3D) protein-ligand complexes is critical for drug repositioning and virtual drug screening. Existing approaches usually transform a 3D protein-ligand complex to a two-dimensional (2D) graph, and then use graph neural networks (GNNs) to predict its binding affinity. However, the node and edge features of the 2D graph are extracted based on invariant local coordinate systems of the 3D complex. As a result, these approaches can not fully learn the global information of the complex, such as the physical symmetry and the topological information of bonds. To address these issues, we propose a novel Equivariant Line Graph Network (ELGN) for binding affinity prediction of 3D protein-ligand complexes. The proposed ELGN firstly adds a super node to the 3D complex, and then builds a line graph based on the 3D complex. After that, ELGN uses a new E(3)-equivariant network layer to pass the messages between nodes and edges based on the global coordinate system of the 3D complex. Experimental results on two real datasets demonstrate the effectiveness of ELGN over several state-of-the-art baselines.

Epilepsy is a prevalent neurological disorder, yet its underlying mechanisms remain incompletely understood. Accumulated studies have indicated that epilepsy is characterized by abnormal neural circuits. Understanding the circuit mechanisms is crucial for comprehending the pathogenesis of epilepsy. With advances in tracing and modulating tools for neural circuits, some epileptic circuits have been uncovered. This comprehensive review focuses on the circuit mechanisms underlying epilepsy in various neuronal subtypes, elucidating their distinct roles. Epileptic seizures are primarily characterized by the hyperactivity of glutamatergic neurons and inhibition of GABAergic neurons. However, specific activated GABAergic neurons and suppressed glutamatergic neurons exacerbate epilepsy through preferentially regulating the activity of GABAergic neurons within epileptic circuits. Distinct subtypes of GABAergic neurons contribute differently to epileptic activities, potentially due to their diverse connection patterns. Moreover, identical GABAergic neurons may assume distinct roles in different stages of epilepsy. Both GABAergic neurons and glutamatergic neurons with long-range projecting fibers innervate multiple nuclei; nevertheless, not all of these circuits contribute to epileptic activities. Epileptic circuits originating from the same nuclei may display diverse contributions to epileptic activities, and certain glutamatergic circuits from the same nuclei may even exert opposing effects on epilepsy. Neuromodulatory neurons, including cholinergic, serotonergic, dopaminergic, and noradrenergic neurons, are also implicated in epilepsy, although the underlying circuit mechanisms remain poorly understood. These studies suggest that epileptic nuclei establish intricate connections through cell-type-specific circuits and play pivotal roles in epilepsy. However, there are still limitations in knowledge and methods, and further understanding of epileptic circuits is crucial, particularly in the context of refractory epilepsy.

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.

We present JueWu-SL, the first supervised-learning-based artificial intelligence (AI) program that achieves human-level performance in playing multiplayer online battle arena (MOBA) games. Unlike prior attempts, we integrate the macro-strategy and the micromanagement of MOBA-game-playing into neural networks in a supervised and end-to-end manner. Tested on Honor of Kings, the most popular MOBA at present, our AI performs competitively at the level of High King players in standard 5v5 games. [ABSTRACT FROM AUTHOR]

Elucidating connectivity and functionality at the whole-brain level is one of the most challenging research goals in neuroscience. Various whole-brain optical imaging technologies with submicron lateral resolution have been developed to reveal the fine structures of brain-wide neural and vascular networks at the mesoscopic level. Among them, micro-optical sectioning tomography (MOST) is attracting increasing attention, as a variety of technological variations and solutions tailored toward different biological applications have been optimized. Here, we summarize the recent development of MOST technology in whole-brain imaging and anticipate future improvements.

One critical deficiency of traditional online kernel learning methods is their unbounded and growing number of support vectors in the online learning process, making them inefficient and non-scalable for large-scale applications. Recent studies on scalable online kernel learning have attempted to overcome this shortcoming, e.g., by imposing a constant budget on the number of support vectors. Although they attempt to bound the number of support vectors at each online learning iteration, most of them fail to bound the number of support vectors for the final output hypothesis, which is often obtained by averaging the series of hypotheses over all the iterations. In this article, we propose a novel framework for bounded online kernel methods, named “Sparse Passive-Aggressive (SPA)” learning, which is able to yield a final output kernel-based hypothesis with a bounded number of support vectors. Unlike the common budget maintenance strategy used by many existing budget online kernel learning approaches, the idea of our approach is to attain the bounded number of support vectors using an efficient stochastic sampling strategy that samples an incoming training example as a new support vector with a probability proportional to its loss suffered. We theoretically prove that SPA achieves an optimal mistake bound in expectation, and we empirically show that it outperforms various budget online kernel learning algorithms. Finally, in addition to general online kernel learning tasks, we also apply SPA to derive bounded online multiple-kernel learning algorithms, which can significantly improve the scalability of traditional Online Multiple-Kernel Classification (OMKC) algorithms while achieving satisfactory learning accuracy as compared with the existing unbounded OMKC algorithms.

BACKGROUND:Immunotherapy using dendritic cell (DC) vaccine has the potential to overcome the bottleneck of cancer therapy. METHODS:We engineered Lewis lung cancer cells (LLCs) and bone marrow–derived DCs to express tumor-associated antigen (TAA) ovalbumin (OVA) via lentiviral vector plasmid encoding OVAgene. We then tested the antitumor effect of modified DCs both in vitroand in vivo. RESULTS:The results demonstrated that in vitromodified DCs could dramatically enhance T-cell proliferation (P<.01) and killing of LLCs than control groups (P<.05). Moreover, modified DCs could reduce tumor size and prolong the survival of LLC tumor-bearing mice than control groups (P<.01 and P<.01, respectively). Mechanistically, modified DCs demonstrated enhanced homing to T-cell–rich compartments and triggered more naive T cells to become cytotoxic T lymphocytes, which exhibited significant infiltration into the tumors. Interestingly, modified DCs also markedly reduced tumor cells harboring stem cell markers in mice (P<.05), suggesting the potential role on cancer stem-like cells. CONCLUSION:These findings suggested that DCs bioengineered with TAA could enhance antitumor effect and therefore represent a novel anticancer strategy that is worth further exploration.

Online learning plays an important role in many big data mining problems because of its high efficiency and scalability. In the literature, many online learning algorithms using gradient information have been applied to solve online classification problems. Recently, more effective second-order algorithms have been proposed, where the correlation between the features is utilized to improve the learning efficiency. Among them, Confidence-Weighted (CW) learning algorithms are very effective, which assume that the classification model is drawn from a Gaussian distribution, which enables the model to be effectively updated with the second-order information of the data stream. Despite being studied actively, these CW algorithms cannot handle nonseparable datasets and noisy datasets very well. In this article, we propose a family of Soft Confidence-Weighted (SCW) learning algorithms for both binary classification and multiclass classification tasks, which is the first family of online classification algorithms that enjoys four salient properties simultaneously: (1) large margin training, (2) confidence weighting, (3) capability to handle nonseparable data, and (4) adaptive margin. Our experimental results show that the proposed SCW algorithms significantly outperform the original CW algorithm. When comparing with a variety of state-of-the-art algorithms (including AROW, NAROW, and NHERD), we found that SCW in general achieves better or at least comparable predictive performance, but enjoys considerably better efficiency advantage (i.e., using a smaller number of updates and lower time cost). To facilitate future research, we release all the datasets and source code to the public at http://libol.stevenhoi.org/.

Three-dimensional high-resolution anorectal manometry (3DHRAM) is a new technique that can explore anorectal disorders and provide interesting topographic data for the diagnosis of pelvic floor disorders such as paradoxical puborectalis syndrome (PPS). Our object was to evaluate whether 3DHRAM can reliably diagnose PPS already diagnosed with X-ray defaecography, which is considered to be the gold standard.All patients being tested in our department for dyschezia by 3D-HRAM and X-ray defecography were eligible for the study. The 3DHRAM results were compared with X-ray defecography. The sensitivity, specificity, and positive and negative predictive values were calculated for various 3DHRAM criteria to propose a diagnostic strategy for PPS.Twenty-three patients presented with PPS on X-ray defaecography. On 3DHRAM, according to our diagnostic strategy, the kappa value was 0.706, with a positive predictive value of 71.88% [95% CI, 53.02–85.60], a specificity of 80.43% [95% CI, 65.62–90.13], a sensibility of 95.83% [95% CI, 76.98–99.78], and area under curve value was 0.922.In this study, 3DHRAM was used to diagnose PPS with the same degree of reliability as X-ray defaecography, and we confirmed its use in the diagnosis of pelvic floor disorders. Further studies will be necessary to define classifications for these new anatomic data from 3DHRAM.

We have previously shown that tyrosine phosphorylation of the actin-regulatory protein villin is accompanied by the redistribution of phosphorylated villin and a concomitant decrease in the F-actin content of intestinal epithelial cells. The temporal and spatial correlation of these two events suggested that tyrosine phosphorylation of villin may be involved in the rearrangement of the microvillar cytoskeleton. This hypothesis was investigated by analyzing the effects of tyrosine phosphorylation of villin on the kinetics of actin polymerization by reconstituting in vitrothe tyrosine phosphorylation of villin and its association with actin. Full-length recombinant human villin was phosphorylated in vitroby expression in the TKX1-competent cells that carry an inducible tyrosine kinase gene. The actin-binding properties of villin were examined using a co-sedimentation assay. Phosphorylation of villin did not change the stoichiometry (1:2) but decreased the binding affinity (4.4 μmfor unphosphorylatedversus0.6 μmfor phosphorylated) of villin for actin. Using a pyrene-actin-based fluorescence assay, we demonstrated that tyrosine phosphorylation had a negative effect on actin nucleation by villin. In contrast, tyrosine phosphorylation enhanced actin severing by villin. Electron microscopic analysis showed complementary morphological changes. Phosphorylation inhibited the actin bundling and enhanced the actin severing functions of villin. Taken together our data show that tyrosine phosphorylation of villin decreases the amount of villin bound to actin filaments, inhibits the actin-polymerizing properties of villin, and promotes the actin-depolymerizing functions instead. These observations suggest a role for tyrosine phosphorylation in modulating the microvillar cytoskeleton in vivoby villin in response to specific physiological stimuli.