Backpropagation (BP) is the most successful and widely used algorithm in deep learning. However, the computations required by BP are challenging to reconcile with known neurobiology. This difficulty has stimulated int...
ISBN:
(纸本)9781713871088
Backpropagation (BP) is the most successful and widely used algorithm in deep learning. However, the computations required by BP are challenging to reconcile with known neurobiology. This difficulty has stimulated interest in more biologically plausible alternatives to BP. One such algorithm is the inference learning algorithm (IL). IL trains predictive coding models of neural circuits and has achieved equal performance to BP on supervised and auto-associative tasks. In contrast to BP, however, the mathematical foundations of IL are not well-understood. Here, we develop a novel theoretical framework for IL. Our main result is that IL closely approximates an optimization method known as implicit stochastic gradient descent (implicit SGD), which is distinct from the explicit SGD implemented by BP. Our results further show how the standard implementation of IL can be altered to better approximate implicit SGD. Our novel implementation considerably improves the stability of IL across learning rates, which is consistent with our theory, as a key property of implicit SGD is its stability. We provide extensive simulation results that further support our theoretical interpretations and find IL achieves quicker convergence when trained with mini-batch size one while performing competitively with BP for larger mini-batches when combined with Adam.
Multi-omics data clustering aims at assigning elements to their respective categories via exploiting the complementary information among multi-omics samples, with the widely application on data mining, bioinformatics,...
详细信息
Virtual reality technology provides a strong sense of immersion and interactivity. It is widely used in the fields of anxiety relief, fear therapy, and depression regulation. However, objectively evaluating the emotio...
详细信息
Eye-tracking holds numerous promises for improving the mixed reality experience. While eye-tracking devices are capable of accurate gaze mapping on 2D surfaces, depth estimation of gaze points remains a challenging pr...
详细信息
Multimode fibers (MMFs) have recently reemerged as attractive avenues for nonlinear effects due to their high-dimensional spatiotemporal nonlinear dynamics and scalability for high power. High-brightness MMF sources w...
详细信息
Despite the tremendous success of automatic speech recognition (ASR) with the introduction of deep learning, its performance is still unsatisfactory in many real-world multi-talker scenarios. Speaker separation excels...
详细信息
The achievement of emotional functions relies on the interactions of various functional systems of the human brain. Numerous studies tried to explore the mechanism of the emotions based on the functional connectivity....
详细信息
ISBN:
(数字)9798350322996
ISBN:
(纸本)9798350323009
The achievement of emotional functions relies on the interactions of various functional systems of the human brain. Numerous studies tried to explore the mechanism of the emotions based on the functional connectivity. However, the contribution of the transient and stable patterns of brain communication in brain emotions was still unclear. The recently proposed activation network framework assumed the activity of functional connectivity (AFC) and background of functional connectivity (BFC) to respectively represent transient and stable patterns of functional connectivity. In this paper, we employed the activation network framework to SEED-IV dataset to achieve the emotion recognition and evaluate the performance of the transient and stable patterns in emotional activities. The top 100 critical connections of each subject were extracted by a data-driven feature selection strategy. The critical connections across all subjects of both AFC and BFC suggested the importance of Gamma band in emotion recognition. Especially, the AFC and BFC showed the different communication modes during the emotions. Finally, the subject-independent classification was employed on each subject’s critical connections to achieve the emotion recognition. The BFC showed the best classification accuracy of 78.71% ± 1.73% (mean ± std). The findings demonstrated that human emotions were mostly influenced by the consistent brain communication patterns. The findings of this investigation offer a new insight on the studies of human emotion.
brain-computer interface (BCI) has garnered the significant attention for their potential in various applications, with event-related potential (ERP) performing a considerable role in BCI systems. This paper introduce...
brain-computer interface (BCI) has garnered the significant attention for their potential in various applications, with event-related potential (ERP) performing a considerable role in BCI systems. This paper introduces a novel Distributed Inference System tailored for detecting task-wise single-trial ERPs in a stream of satellite images. Unlike traditional methodologies that employ a single model for target detection, our system utilizes multiple models, each optimized for specific tasks, ensuring enhanced performance across varying image transition times and target onset times. Our experiments, conducted on four participants, employed two paradigms: the Normal paradigm and an AI paradigm with bounding boxes. Results indicate that our proposed system outperforms the conventional methods in both paradigms, achieving the highest $F_{\beta}$ scores. Furthermore, including bounding boxes in the AI paradigm significantly improved target recognition. This study underscores the potential of our Distributed Inference System in advancing the field of ERP detection in satellite image streams.
Several Convolutional Deep Learning models have been proposed to classify the cognitive states utilizing several neuro-imaging domains. These models have achieved significant results, but they are heavily designed wit...
详细信息
Varying dynamics parameters in simulation is a popular Domain Randomization (DR) approach for overcoming the reality gap in Reinforcement Learning (RL). Nevertheless, DR heavily hinges on the choice of the sampling di...
详细信息
暂无评论