Smart home hubs have significantly improved everyday home life by serving as central control units that connect and manage various devices, such as lights, door locks, curtains, and cameras. However, the limited CPU and memory resources of these hubs hinder the execution of multiple intelligent tasks, such as human activity recognition, image classification, and speech recognition. To fully utilize these resources, we proposeS-HUB, a scalable deep neural network (DNN) fusion framework for smart home hubs capable of handling multiple tasks. In the offline phase, we apply DNN pruning, weight virtualization, and re-fusion to create a unified model that dynamically scales across tasks. In the online phase, we design a DNN scheduling optimizer to achieve optimal multitask inference while adhering to resource constraints. Finally, comparative experiments and evaluations in smart home scenarios are conducted to assess the performance ofS-HUBacross various tasks, including speech recognition, object detection, gesture recognition, food recognition, and fall detection. Experimental results show that compared to two state-of-the-art baselines,S-HUBachieves an average task processing time of 3 seconds (at least 14% faster) under accuracy constraints, and an average accuracy loss rate of 4.21% (at least 68% lower) under latency constraints. In unconstrained scenarios, it consistently delivers the best overall performance (2.99 seconds and 2.94% loss), demonstrating the scalability and effectiveness of our fusion method in handling performance-resource trade-offs across tasks.

Edge-cloud collaborative learning emerges as a promising paradigm for adapting pre-trained deep neural network (DNN) models to the ever-changing edge data environments and specific downstream tasks. However, the heterogeneity of edge devices and unlabeled local data hinder the effectiveness of existing collaborative learning approaches. To address the above issues, we proposeACL, a novel adaptive edge-cloud collaborative learning paradigm for heterogeneous devices with unlabeled local data. InACL, we first useFedNAS, a neural architecture search algorithm designed for collaborative learning to generate a customized model on each participating device, and then a lightweight semi-supervised collaborative learning frameworkHSSCLis used to fine-tune the pre-trained DNN model. Compared with the SOTA collaborative learning approaches,ACLachieves significant accuracy improvement, averaging 31.5% for image classification and 15.5% for object detection. Furthermore, it reduces time overhead by 3.1-5.1× and memory overhead by 6.3-12.5×. We will release our models and tools.

Gravitational Search Algorithm (GSA) is a well‐known physics‐based meta‐heuristic algorithm inspired by Newton’s law of universal gravitation and performs well in solving optimization problems. However, when solving high‐dimensional optimization problems, the performance of GSA may deteriorate dramatically due to severe interference of redundant dimensional information in the high‐dimensional space. To solve this problem, this paper proposes a Manifold‐Guided Gravitation Search Algorithm, called MGGSA. First, based on the Isomap, an effective dimension extraction method is designed. In this mechanism, the effective dimension is extracted by comparing the dimension differences of the particles located in the same sorting position both in the original space and the corresponding low‐dimensional manifold space. Then, the gravitational adjustment coefficient is designed, so that the particles can be guided to move in a more appropriate direction by increasing the effect of effective dimension, reducing the interference of redundant dimension on particle motion. The performance of the proposed algorithm is tested on 35 high‐dimensional (dimension is 1000) benchmark functions from CEC2010 and CEC2013, and compared with eleven state‐of‐art meta‐heuristic algorithms, the original GSA and four latest GSA’s variants, as well as three well‐known large‐scale global optimization algorithms. The experimental results demonstrate that MGGSA not only has a fast convergence rate but also has high solution accuracy. Besides, MGGSA is applied to three real‐world application problems, which verifies the effectiveness of MGGSA on practical applications.

Deploying deep learning models to edge devices for low-latency and privacy-preserving applications has become a trend. To adapt to heterogeneous devices and data, it is significant to generate customized models. However, existing model adaptation approaches require edge devices to make interactions (collecting hardware information or local data) with the cloud, which raises privacy concerns, increases communication costs, and burdens the cloud. By contrast, we proposeCamoNet, a universal on-device model adaptation framework with zero interaction between devices and the cloud. InCamoNet, a lightweight on-device neural architecture search module is utilized to quickly generate a customized model for subsequent on-device training, followed by an on-device contrastive transfer learning module to effectively leverage unlabeled data for fine-tuning the customized model. Extensive experimental results show thatCamoNetcan effectively run on various edge devices. Compared with the SOTA model adaptation approaches,CamoNetachieves significant accuracy improvement by 25.2% on average for image classification, 10.1% on average for object detection, and reduces the training memory by 4.8-11.4×. We will open-source our models and tools for edge AI developers.

In recent years, many advances have been made in the development of molecular scale devices. Experimental data shows that these devices have potential for use in both memory and logic. This article describes the challenges faced in building crossbar array-based molecular memory and develops a methodology to optimize molecular scale architectures based on experimental device data taken at room temperature. In particular, issues in reading and writing such as memory using CMOS are discussed, and a solution is introduced for easily reading device conductivity states (typically characterized by very small currents). Additionally, a metric is derived to determine the voltages for writing to the crossbar array. The proposed memory design is also simulated with consideration to device parameter variations. Thus, the results presented here shed light on important design choices to be made at multiple abstraction levels, from devices to architectures. Simulation results, incorporating experimental device data, are presented using Cadence Spectre.

In recent years many advances have been made in the development of molecular scale devices. Experimental data shows that these devices have potential for use in both memory and logic. This paper describes the challenges faced in building crossbar array based molecular memory, and develops a methodology to optimize molecular scale architectures based on experimental device data taken at room temperature. In particular, we discuss reading and writing such memory using CMOS and compiling a solution for easily reading device conductivity states (typically characterized by very small currents). Additionally, a metric is derived to determine the voltages for writing to the crossbar array. Simulation results, incorporating experimental device data, are presented using Cadence Spectre.