Reducing Parameters of Neural Networks via Recursive Tensor Approximation

Large-scale neural networks have attracted much attention for surprising results in various cognitive tasks such as object detection and image classification. However, the large number of weight parameters in the complex networks can be problematic when the models are deployed to embedded systems. In addition, the problems are exacerbated in emerging neuromorphic computers, where each weight parameter is stored within a synapse, the primary computational resource of the bio-inspired computers. We describe an effective way of reducing the parameters by a recursive tensor factorization method. Applying the singular value decomposition in a recursive manner decomposes a tensor that represents the weight parameters. Then, the tensor is approximated by algorithms minimizing the approximation error and the number of parameters. This process factorizes a given network, yielding a deeper, less dense, and weight-shared network with good initial weights, which can be fine-tuned by gradient descent.

Development of Sustainable Energy Use with Attention to Fruitful Policy

There are different energy approaches around the world to the development of sustainable energy systems. In this regard, the role of governments, local governments, and people in the development and use of sustainable energy is remarkable. This research, concerning the present epistemic and normative differences, aims to investigate the societal debate on citizen inclusion, local and national attempts to develop clear procedures and guidelines in the transition to sustainable energy use in different countries. Existing theories, subjectivities, and policy implications for different countries are first carefully analyzed. Based on theories, evidence, and policy implications, the behavioural insights for sustainable energy use are then examined. The results show that national governments should never ignore the psychology and behaviour of people, especially in terms of economic behaviour, performance applicable and knowledge of local governments and people in sustainable energy development. Channels of communication between local, people, and national governments, can make a robust shared network and implement simple policies such as increasing their authority. They can also encourage and build capacity through the training, support, trust and knowledge capacity of local governments and people to move toward sustainable energy development. Therefore, focusing on government and maintaining national authority should be departed from any approaches that local government and the public should be constrained as minor actors in sustainable energy governance networks. This work demonstrates that local governments can develop sustainable energy. Moreover, national governments can overcome issues and further control sustainable energy public policy goals under difficult national political conditions.

Loneliness: An Immunometabolic Syndrome

Loneliness has been defined as an agonizing encounter, experienced when the need for human intimacy is not met adequately, or when a person’s social network does not match their preference, either in number or attributes. This definition helps us realize that the cause of loneliness is not merely being alone, but rather not being in the company we desire. With loneliness being introduced as a measurable, distinct psychological experience, it has been found to be associated with poor health behaviors, heightened stress response, and inadequate physiological repairing activity. With these three major pathways of pathogenesis, loneliness can do much harm; as it impacts both immune and metabolic regulation, altering the levels of inflammatory cytokines, growth factors, acute-phase reactants, chemokines, immunoglobulins, antibody response against viruses and vaccines, and immune cell activity; and affecting stress circuitry, glycemic control, lipid metabolism, body composition, metabolic syndrome, cardiovascular function, cognitive function and mental health, respectively. Taken together, there are too many immunologic and metabolic manifestations associated with the construct of loneliness, and with previous literature showcasing loneliness as a distinct psychological experience and a health determinant, we propose that loneliness, in and of itself, is not just a psychosocial phenomenon. It is also an all-encompassing complex of systemic alterations that occur with it, expanding it into a syndrome of events, linked through a shared network of immunometabolic pathology. This review aims to portray a detailed picture of loneliness as an “immunometabolic syndrome”, with its multifaceted pathology.

Transmembrane Batten disease proteins interact with a shared network of vesicle sorting proteins to regulate synaptic composition and function.

Batten disease is unique among lysosomal storage disorders for the early and profound manifestation in the central nervous system, but little is known regarding potential neuron-specific roles for the disease-associated proteins. We demonstrate substantial overlap in the protein interactomes of three transmembrane Batten proteins (CLN3, CLN6, and CLN8), and that their absence leads to synaptic depletion of key partners (i.e. SNAREs and tethers) and aberrant synaptic SNARE dynamics in vivo, demonstrating a novel shared etiology.

ViTT: Vision Transformer Tracker

This paper presents a new model for multi-object tracking (MOT) with a transformer. MOT is a spatiotemporal correlation task among interest objects and one of the crucial technologies of multi-unmanned aerial vehicles (Multi-UAV). The transformer is a self-attentional codec architecture that has been successfully used in natural language processing and is emerging in computer vision. This study proposes the Vision Transformer Tracker (ViTT), which uses a transformer encoder as the backbone and takes images directly as input. Compared with convolution networks, it can model global context at every encoder layer from the beginning, which addresses the challenges of occlusion and complex scenarios. The model simultaneously outputs object locations and corresponding appearance embeddings in a shared network through multi-task learning. Our work demonstrates the superiority and effectiveness of transformer-based networks in complex computer vision tasks and paves the way for applying the pure transformer in MOT. We evaluated the proposed model on the MOT16 dataset, achieving 65.7% MOTA, and obtained a competitive result compared with other typical multi-object trackers.

Neural correlates of ingroup bias for prosociality in rats

Prosocial behavior, in particular helping others in need, occurs preferentially in response to distress of one’s own group members. In order to explore the neural mechanisms promoting mammalian helping behavior, a discovery-based approach was used here to identify brain-wide activity correlated with helping behavior in rats. Demonstrating social selectivity, rats helped others of their strain (‘ingroup’), but not rats of an unfamiliar strain (‘outgroup’), by releasing them from a restrainer. Analysis of brain-wide neural activity via quantification of the early-immediate gene c-Fos identified a shared network, including frontal and insular cortices, that was active in the helping test irrespective of group membership. In contrast, the striatum was selectively active for ingroup members, and activity in the nucleus accumbens, a central network hub, correlated with helping. In vivo calcium imaging showed accumbens activity when rats approached a trapped ingroup member, and retrograde tracing identified a subpopulation of accumbens-projecting cells that was correlated with helping. These findings demonstrate that motivation and reward networks are associated with helping an ingroup member and provide the first description of neural correlates of ingroup bias in rodents.

Dynamic Event-triggered Control and Estimation: A Survey

The efficient utilization of computation and communication resources became a critical design issue in a wide range of networked systems due to the finite computation and processing capabilities of system components (e.g., sensor, controller) and shared network bandwidth. Event-triggered mechanisms (ETMs) are regarded as a major paradigm shift in resource-constrained applications compared to the classical time-triggered mechanisms, which allows a trade-off to be achieved between desired control/estimation performance and improved resource efficiency. In recent years, dynamic event-triggered mechanisms (DETMs) are emerging as a promising enabler to fulfill more resource-efficient and flexible design requirements. This paper provides a comprehensive review of the latest developments in dynamic event-triggered control and estimation for networked systems. Firstly, a unified event-triggered control and estimation framework is established, which empowers several fundamental issues associated with the construction and implementation of the desired ETM and controller/estimator to be systematically investigated. Secondly, the motivations of DETMs and their main features and benefits are outlined. Then, two typical classes of DETMs based on auxiliary dynamic variables (ADVs) and dynamic threshold parameters (DTPs) are elaborated. In addition, the main techniques of constructing ADVs and DTPs are classified, and their corresponding analysis and design methods are discussed. Furthermore, three application examples are provided to evaluate different ETMs and verify how and under what conditions DETMs are superior to their static and periodic counterparts. Finally, several challenging issues are envisioned to direct the future research.

Functional coupling between target selection and acquisition in the superior colliculus

To survive in unpredictable environments, animals must continuously evaluate their surroundings for behavioral targets, such as food and shelter, and direct their movements to acquire those targets. Although the ability to accurately select and acquire spatial targets depends on a shared network of brain regions, how these processes are linked by neural circuits remains unknown. The superior colliculus (SC) mediates the selection of spatial targets and remains active during orienting movements to acquire targets, which suggests the underexamined possibility that common SC circuits underie both selection and acquisition processes. Here, we test the hypothesis that SC functional circuitry couples target selection and acquisition using a default motor plan generated by selection-related neuronal activity. Single-unit recordings from intermediate and deep layer SC neurons in male mice performing a spatial choice task demonstrated that choice-predictive neurons, including optogenetically identified GABAergic SC neurons whose activity was causally related to target selection, exhibit increased activity during movement to the target. By strategically recording from both rostral and caudal SC neurons, we also revealed an overall caudal-to-rostral shift in activity as targets were acquired. Finally, we used an attractor model to examine how target selection activity in the SC could generate a rostral shift in activity during target acquisition using only intrinsic SC circuitry. Overall, our results suggest a functional coupling between SC circuits that underlie target selection and acquisition, elucidating a key mechanism for goal-directed behavior.