The success of deep learning models heavily depends on high-quality labeled data, yet noisy labels are inevitable in large-scale datasets. Existing methods often suffer from confirmation bias and overlook the informative value of hard but clean samples. To address these challenges, we propose Local Semantics Refinement of Adaptive Representations (LFDA), a novel framework that adaptively refines label quality by leveraging local feature consistency and representation alignment. LFDA introduces a Local Consistency Score module that evaluates the similarity among local samples in the latent space to distinguish clean from noisy labels. In addition, a confidence neighborhood is further constructed to provide local reference guidance, enabling more accurate identification and correction of noisy instances. To enhance semantic reliability, LFDA integrates a Reliability-Aware Representation Alignment (RRA) module that aligns high-confidence sample representations to implicitly refine low-confidence instances via soft supervision. Extensive experiments on both synthetic and real-world noisy datasets demonstrate that LFDA consistently outperforms state-of-the-art label noise learning methods. The results confirm its good robustness, generalization ability, and effectiveness in handling diverse and complex noise conditions.

Few-Shot Medical Image Segmentation (FSMIS) aims to segment novel object classes in medical images using only minimal annotated examples, addressing the critical challenges of data scarcity and domain shifts prevalent in medical imaging. While Diffusion Models (DM) excel in visual tasks, their potential for FSMIS remains largely unexplored. We propose that the rich visual priors learned by large-scale DMs offer a powerful foundation for a more robust and data-efficient segmentation approach. In this paper, we introduce SD-FSMIS, a novel framework designed to effectively adapt the powerful pre-trained Stable Diffusion (SD) model for the FSMIS task. Our approach repurposes its conditional generative architecture by introducing two key components: a Support-Query Interaction (SQI) and a Visual-to-Textual Condition Translator (VTCT). Specifically, SQI provides a straightforward yet powerful means of adapting SD to the FSMIS paradigm. The VTCT module translates visual cues from the support set into an implicit textual embedding that guides the diffusion model, enabling precise conditioning of the generation process. Extensive experiments demonstrate that SD-FSMIS achieves competitive results compared to state-of-the-art methods in standard settings. Surprisingly, it also demonstrated excellent generalization ability in more challenging cross-domain scenarios. These findings highlight the immense potential of adapting large-scale generative models to advance data-efficient and robust medical image segmentation.

Open-World Object Detection (OWOD) requires detectors to identify previously unseen objects as unknown and incrementally incorporate them into the set of known categories, while preserving previously acquired knowledge. Existing frameworks rely heavily on exemplar replay to mitigate catastrophic forgetting, but in some real applications, storing raw data conflicts with data access restrictions and leads to data exposure risks, while incurring significant memory overhead. In this paper, we propose REAL-OW, a novel rehearsal-free framework that decouples incremental knowledge through a collaborative adapter architecture based on Low-Rank Adaptation (LoRA). Specifically, we deploy General Adapters (GAs) in the backbone to enable the significance-aware refinement of cross-task universal representations, while Specific Adapters (SAs) in the decoder provide orthogonal storage for task-specific expertise. To resolve representation drift in objectness modeling under rehearsal-free constraints, we introduce Dual-Stage Objectness Modeling (DSOM), which alternates between feature aggregation and boundary consolidation to stabilize objectness distributions while maintaining the separation between known and unknown categories. Furthermore, DSOM is supported by a Calibrated Gaussian Negative Log-Likelihood (CG-NLL) distance tailored for the dispersed feature distributions inherent in rehearsal-free settings. Extensive evaluations demonstrate that REAL-OW achieves state-of-the-art performance, surpassing existing exemplar replay methods in both detection precision and unknown discovery. Our approach establishes a new baseline for rehearsal-free OWOD.

In recent years, significant advancements in Spiking Neural Networks (SNNs) have driven increasing interest in applying Neural Architecture Search (NAS) to develop more advanced SNN models. However, existing NAS methods in SNNs often adopt structures from mature Deep Neural Networks (DNNs), which are predominantly feedforward and neglect neuronal diversity. Additionally, SNNs inherently require simulated timesteps during training, resulting in higher memory consumption and latency compared to Artificial Neural Networks (ANNs). To address these limitations, we propose Sparsity-Aware Evolutionary Neural Architecture Search (SAE-NAS) to design brain-inspired SNNs. Specifically, we introduce a brain-inspired search space comprising diverse neural circuits that integrate excitatory neurons, inhibitory neurons, and feedback connections. This design simulates biological neural systems by establishing a dynamic excitation-inhibition equilibrium, which enhances information processing capacity while mitigating overactivation through inhibitory neuron regulation. To minimize computational overhead, we present the Sparsity-Aware Evolutionary (SAE) approach, which efficiently discovers high-performance architectures without memory-intensive backpropagation or iterative parameter updates. Furthermore, we propose the Sparsity-Aware Score (SAS), a training-free metric that quantifies network expressivity using only a single forward pass on a small batch. Extensive experiments demonstrate the efficacy of our method, achieving SOTA performance on both static and neuromorphic datasets.

Previous studies on memristive associative neural network circuits have not sufficiently incorporated the effects of spike-rate-dependent plasticity (SRDP) together with long-term potentiation (LTP) and long-term depression (LTD) mechanisms. To address this limitation, based on the SRDP rule, this paper presents the design of a full-function memristive associative memory neural network circuit. The circuit adopts an improved leaky integrate-and-fire (LIF) neuron to generate graded pulse signals, which are subsequently classified into three frequency ranges by a frequency recognition module. Based on the identified frequency, the long-term potentiation or depression module is activated to execute SRDP operations: High-frequency signals induce neurons to produce LTP, which increases the rate of learning and decreases the rate of forgetting; Medium-frequency signals induce neurons to produce LTD, which decreases the rate of learning and increases the rate of forgetting; Low-frequency signals do not affect the rate of change of synaptic weights in associative memory. The updated synaptic weights are then routed to an associative memory module, where the circuit implements learning and forgetting (ring-only and food-only patterns). The circuit further incorporates a consolidation learning module to counteract natural forgetting and stabilize synaptic weights for long-term memory retention. By integrating these modules, the circuit simulates learning, forgetting, and long-term memory formation in memristive associative neural networks. The functionality and effectiveness of the circuit is verified by PSPICE simulation.

Address event representation (AER) object recognition task has attracted extensive attention in neuromorphic vision processing. The spike-based and event-driven computation inherent in the spiking neural network (SNN) provides an energy-saving solution for AER object recognition. However, SNN with spike timing dependent plasticity (STDP) learning rule has not achieved satisfying AER object recognition performance. This work proposes an SNN-based encoder-decoder model to improve the recognition performance of AER objects. An STDP-based locally connected spiking neural network (LC-SNN) is proposed as an encoder to extract rich spatiotemporal features from AER event flows more flexibly. After the encoder extracts and learns primary features, we propose a fully connected spiking neural network (FC-SNN) based on the reward-modulated spike-timing-dependent plasticity (R-STDP) learning rule as a decoder to learn higher-level features for classification. In addition, we improved the winner-take-all (WTA) mechanisms and R-STDP learning rule in the decoder based on the reward and punish decision, enabling the network to perform better. The experiments are performed on the N-MNIST, MNIST-DVS, and the dynamic vision sensor (DVS) gesture datasets, improving the accuracy of the best existing plasticity-based SNN by 0.19%, 0.27%, and 1.35%, respectively.

While backpropagation effectively trains models, it presents challenges related to bio-plausibility, resulting in high memory demands and limited parallelism. Recently, Hinton (2022) proposed the Forward-Forward (FF) algorithm for high-parallel local updates. FF leverages squared sums as the local update target, termed goodness, and employs L2 normalization to decouple goodness and extract new features. However, this design encounters issues with feature scaling and deactivated neurons, limiting its application mainly to shallow networks. This paper proposes a novel goodness design utilizing **layer normalization** and **mean goodness** to overcome these challenges, demonstrating performance improvements even in 17-layer CNNs. Experiments on CIFAR-10, MNIST, and Fashion-MNIST show significant advantages over existing FF-based algorithms, highlighting the potential of FF in deep models. Furthermore, the model parallel strategy is proposed to achieve highly efficient training based on the property of local updates.

Sign Language Recognition (SLR) is an increasingly popular research topic due to its extensive potential applications, such as education, healthcare, emergency response, and social interaction. Sign language is a complex and dynamic language comprising hand gestures, facial expressions, and body motions. This high level of variability poses a significant obstacle for SLR tasks, which must accurately identify and respond to numerous gestures. To address these challenges, an end-to-end skeleton-based multi-feature multi-stream multi-level information sharing network (three multi information sharing network (TMS-Net)) is proposed. Specifically, in order to input more rich information to TMS-Net, we use joint feature pair with global features, bone feature pair with local features, and angle feature pair with scale invariance. In terms of network structure, to efficiently extract multiple features from inputs, we build a multi-stream structure and design a multi-level information sharing mechanism based on this structure to ensure the full utilization of skeleton feature information. From the experiment results of the WLASL-2000 dataset (56.4%), AUTSL dataset (96.62%) and MSASL(65.13%), TMS-Net surpass the state-of-the-art (SOTA) methods with single modality as input. In addition, a SLR-based human–robot interaction (HRI) experiment using our proposed TMS-Net is conducted, which proves the practical performance of the TMS-Net.

Convolutional neural networks (CNNs) have achieved significant improvement for the task of facial expression recognition. However, current training still suffers from the inconsistent learning intensities among different layers, i.e., the feature representations in the shallow layers are not sufficiently learned compared with those in deep layers. To this end, this work proposes a contrastive learning framework to align the feature semantics of shallow and deep layers, followed by an attention module for representing the multi-scale features in the weight-adaptive manner. The proposed algorithm has three main merits. First, the learning intensity, defined as the magnitude of the backpropagation gradient, of the features on the shallow layer is enhanced by cross-layer contrastive learning. Second, the latent semantics in the shallow-layer and deep-layer features are explored and aligned in the contrastive learning, and thus the fine-grained characteristics of expressions can be taken into account for the feature representation learning. Third, by integrating the multi-scale features from multiple layers with an attention module, our algorithm achieved the state-of-the-art performances, i.e. 92.21%, 89.50%, 62.82%, on three in-the-wild expression databases, i.e. RAF-DB, FERPlus, SFEW, and the second best performance, i.e. 65.29% on AffectNet dataset. Our codes will be made publicly available.

As a possible device to further enhance the performance of the hybrid complementary metal oxide semiconductor (CMOS) technology in the hardware, the memristor has attracted widespread attention in implementing efficient and compact deep learning (DL) systems. In this study, an automatic learning rate tuning method for memristive DL systems is presented. Memristive devices are utilized to adjust the adaptive learning rate in deep neural networks (DNNs). The speed of the learning rate adaptation process is fast at first and then becomes slow, which consist of the memristance or conductance adjustment process of the memristors. As a result, no manual tuning of learning rates is required in the adaptive back propagation (BP) algorithm. While cycle-to-cycle and device-to-device variations could be a significant issue in memristive DL systems, the proposed method appears robust to noisy gradients, various architectures, and different datasets. Moreover, fuzzy control methods for adaptive learning are presented for pattern recognition, such that the over-fitting issue can be well addressed. To our best knowledge, this is the first memristive DL system using an adaptive learning rate for image recognition. Another highlight of the presented memristive adaptive DL system is that quantized neural network architecture is utilized, and there is therefore a significant increase in the training efficiency, without the loss of testing accuracy.

Few-shot semantic segmentation (FSS) endeavors to segment unseen classes with only a few labeled samples. Current FSS methods are commonly built on the assumption that their training and application scenarios share similar domains, and their performances degrade significantly while applied to a distinct domain. To this end, we propose to leverage the cutting-edge foundation model, the Segment Anything Model (SAM), for generalization enhancement. The SAM however performs unsatisfactorily on domains that are distinct from its training data, which primarily comprise natural scene images, and it does not support automatic segmentation of specific semantics due to its interactive prompting mechanism. In our work, we introduce APSeg, a novel auto-prompt network for cross-domain few-shot semantic segmentation (CD-FSS), which is designed to be auto-prompted for guiding cross-domain segmentation. Specifically, we propose a Dual Prototype Anchor Transformation (DPAT) module that fuses pseudo query prototypes extracted based on cycle-consistency with support port prototypes, allowing features to be transformed into a more stable domain-agnostic space. Additionally, a Meta Prompt (MPG) module is introduced to automatically generate prompt embeddings, eliminating the need for manual visual prompts. We build an efficient model which can be applied directly to target domains without fine-tuning. Extensive experiments on four cross-domain datasets show that our model outperforms the state-of-the-art CD-FSS method by 5.24% and 3.10% in average accuracy on 1-shot and 5-shot settings, respectively.

Hair editing is a critical image synthesis task that aims to edit hair color and hairstyle using text descriptions or reference images, while preserving irrelevant attributes (e.g., identity, background, cloth). Many existing methods are based on StyleGAN to address this task. However, due to the limited spatial distribution of StyleGAN, it struggles with multiple hair color editing and facial preservation. Considering the advancements in diffusion models, we utilize Latent Diffusion Models (LDMs) for hairstyle editing. Our approach introduces Multi-stage Hairstyle Blend (MHB), effectively separating control of hair color and hairstyle in diffusion latent space. Additionally, we train a warping module to align the hair color with the target region. To further enhance multi-color hairstyle editing, we fine-tuned a CLIP model using a multi-color hairstyle dataset. Our method not only tackles the complexity of multi-color hairstyles but also addresses the challenge of preserving original colors during diffusion editing. Extensive experiments showcase the superiority of our method in editing multi-color hairstyles while preserving facial attributes given textual descriptions and reference images.

Studying the stability of genetic regulatory networks (GRNs) helps to explore the laws of automatic regulation in organisms. By designing proper controllers, the finite-time stability problem of GRNs with nondifferential discrete time-varying delays is discussed in this brief. Rather than constructing the Lyapunov functional or using the finite-time stability theorem, inequality technique and comparison method are firstly adopted in the analyzation. Then, several sufficient algebraic criteria are derived to ensure the finite-time stability of addressed delayed GRNs. The results are general and contain the exponential stability as a special case. Finally, numerical simulations are provided to illustrate the validness of theoretical results.

The key to controlling the spread of the coronavirus disease 2019 (COVID-19) and reducing mortality is highly dependent on the safe and effective use of vaccines for the general population. Current COVID-19 vaccination practices (intramuscular injection of solution-based vaccines) are limited by heavy reliance on medical professionals, poor compliance, and laborious vaccination recording procedures, resulting in a waste of health resources and low vaccination coverage, etc. In this study, we developed a smart mushroom-inspired imprintable and lightly detachable (MILD) microneedle platform for the effective and convenient delivery of multidose COVID-19 vaccines and decentralized vaccine information storage. The mushroom-like structure allows the MILD system to be easily pressed into the skin and detached from the patch base, acting as a "tattoo" to record the vaccine counts in situ without any storage equipment, offering quick accessibility and effortless readout, saving a great deal of valuable time and energy for both patients and health professionals. After loading inactivated SARS-CoV-2 virus-based vaccines, MILD system induced a high level of antibodies against the SARS-CoV-2 receptor-binding domain (RBD) in vivo without eliciting systemic toxicity and local damage. Collectively, this smart delivery platform serves as a promising carrier to improve COVID-19 vaccination efficacy through its dual capabilities of vaccine delivery and in situ data storage, thus exhibiting great potential for helping to contain the COVID-19 pandemic or a resurgence.

Resistive cross-point memory arrays can be used to construct high-density storage-class memory. However, coupled IR drop and sneak currents cause multidimensional non-uniformity of cell effective voltage in cross-point arrays. The voltage non-uniformity significantly degrades write performance on cross-point memory if only adopting the worst-case write latency at partial dimensions. Furthermore, the non-uniformity of cell effective voltage in cross-point arrays depends on multidimensional dynamic write operation parameters: row, column as well as layer address, the number of selected cells, and the number of half-selected low-resistance state cells. In this article, we aim to improve the write performance by leveraging multidimensional non-uniformity of cell effective voltage. First, we analyze the impact of multidimensional write parameters on effective voltage and write latency. Then, we design the memory array write scheme that measures the write parameters and sets the write latency accordingly. We further analyze the features and effects of interlayer sneak currents and extend the scheme to 3D cross-point memory. The evaluation shows that the proposed memory array write scheme can reduce the memory access latency by 75.6 and 64.1 percent, and improve the system performance by 4.5 times and 3.4 times on average, compared with the baseline and the state-of-the-art approach, respectively.

Resistive memory is promising to be constructed as a high-density storage-class memory. Multilevel cell, access-transistor-free cross-point array structure, and 3-D array integration are three approaches to scale up the density of resistive memory. However, composing the three approaches together strengthens the interactions between array-level and cell-level nonidealities (interconnect resistance-induced IR drop, sneak current, and device variability) of resistive memory arrays during write operations and significantly degrades write performance and reliability. In this article, we analyze the dynamic voltage-dividing effect along a selected write current path in 3-D cross-point memory arrays. We propose a nonideality-tolerant high-density resistive memory (HD-RRAM) architecture, that can weaken the interactions between nonidealities and mitigate their degradation effects on the performance and reliability of array multilevel write operations. HD-RRAM is equipped with a double-transistor array architecture with two-transistor- n-resistor (2TnR) cell organization along pillars to reduce the current driving requirement and the large undesired voltage drop across each vertical pillar access transistor. Moreover, multiside asymmetric bias improves the resistive switching velocity by leveraging current-dividing effects. Variability-aware multilevel state partition reduces the worst-case write error rate by leveraging target state dependency of variability. Proportional-control multilevel state tuning reduces the average number of required write-and-verify iterations by leveraging pulse amplitude dependency of variability. Multilevel cell parallel writing improves the cell-level parallelism by leveraging the pass-through feature of intermediate resistance states. The evaluations show that HD-RRAM reduces both memory access latency and energy consumption over an aggressive baseline.

As a novel nanoscale device, the memristor has elicited widespread interest in implementing compact and efficient neurocomputing systems in the hardware. In this article, fuzzy deep learning systems using fuzzy memristive modeling methods are presented. One key issue in memristive modeling is the device variation issue due to device-to-device and cycle-to-cycle variations. It is very difficult to distinguish the intermediate memristive states in a multilevel memristor. A fuzzy modeling method is therefore proposed to define the memristive states in a dynamic way. In addition, fuzzy deep learning systems are presented for fuzzy pattern recognition such as image recognition. To the authors' best knowledge, this is the first work utilizing memristive fuzzy deep learning systems to realize image recognition. The expected input and output in the memristive fuzzy deep learning systems are refined via a bidirectional fuzzy rule. The effectiveness of the proposed fuzzy methods has been verified with comprehensive methods, such as single-layer neural networks, multi-layer neural networks, convolutional neural networks, and k-nearest neighbor method. Another highlight of the proposed fuzzy deep learning system is that there is a great reduction in memory and a significant increase in the speed for image recognition tasks with the same level of testing accuracy.

Existing deep neural networks (DNNs) are computationally expensive and memory intensive, which hinder their further deployment in novel nanoscale devices and applications with lower memory resources or strict latency requirements. In this paper, a novel approach to accelerate on-chip learning systems using memristive quantized neural networks (M-QNNs) is presented. A real problem of multilevel memristive synaptic weights due to device-to-device (D2D) and cycle-to-cycle (C2C) variations is considered. Different levels of Gaussian noise are added to the memristive model during each adjustment. Another method of using memristors with binary states to build M-QNNs is presented, which suffers from fewer D2D and C2C variations compared with using multilevel memristors. Furthermore, methods of solving the sneak path issues in the memristive crossbar arrays are proposed. The M-QNN approach is evaluated on two image classification datasets, that is, ten-digit number and handwritten images of mixed National Institute of Standards and Technology (MNIST). In addition, input images with different levels of zero-mean Gaussian noise are tested to verify the robustness of the proposed method. Another highlight of the proposed method is that it can significantly reduce computational time and memory during the process of image recognition.

Resistive Random Access Memory (ReRAM) is promising to be employed as high density storage-class memory due to its crossbar array and Triple-Level Cell (TLC) structures. However, TLC crossbar ReRAM suffers from high write latency and energy due to three unique challenges: (1) The crossbar array structure incurs IR drop issues. (2) The TLC structure requires iterative program-and-verify (P&V) procedure. (3) The resistance drift problem needs short interval scrub to avoid uncorrectable soft errors. In this article, to overcome the challenges of TLC crossbar ReRAM, we propose an enhanced low latency and energy efficient TLC crossbar ReRAM architecture, called EnTiered-ReRAM. The proposed EnTiered-ReRAM is composed of four components, including EnTiered-crossbar design, Compression-based Incomplete Data Mapping (CIDM), Compression-based Flip Scheme (CFS), and Compression-based Error Correction Code (CECC). Specifically, based on the observation that our previously proposed Tiered-crossbar design still suffers from large IR drops along bitlines in the far segments due to the long length of bitlines, EnTiered-crossbar partitions each crossbar array into two halves along bitlines, and then splits each bitline of the half crossbar array into the near and far segments by an isolation transistor, which thoroughly mitigates the IR drop issues. Then we use our previously proposed CIDM and CFS in the near and far segments of EnTiered-crossbar arrays to further decreases the write latency and energy. In addition, CECC is deeply coupled with CIDM and CFS. CECC dynamically employs the most appropriate ECC capability according to the remaining space of each cache line after CIDM or CFS encoding, which effectively improves the scrub interval and performance/energy with insignificant space overhead. The evaluation results show that, compared with an aggressive baseline, EnTiered-ReRAM can improve the system performance by 56.3% and reduce the energy consumption by 60.6% on average.

This article provides a review of current development and challenges in brain-inspired computing with memristors. We review the mechanisms of various memristive devices that can mimic synaptic and neuronal functionalities and survey the progress of memristive spiking and artificial neural networks. Different architectures are compared, including spiking neural networks, fully connected artificial neural networks, convolutional neural networks, and Hopfield recurrent neural networks. Challenges and strategies for nanoelectronic brain-inspired computing systems, including device variations, training, and testing algorithms, are also discussed.