Molecular Dynamics (MD) simulation is a fundamental exploring approach for numerous scientific fields, including but not limited to drug discovery, biology, material science, and chemistry. Unfortunately, large-scale MD simulation generally requires long-period processing, even in resource-intensive supercomputers. Parallel computing is the primary methodology to accelerate MD simulation. However, parallel computing has a low theoretical improvable ceiling for MD simulation because the atom statuses of one timestep depend on its previous timestep. To unlock the full potential of MD simulation in drug discovery and biology, we propose an AccelMD framework for reducing processing time and computing costs. AccelMD is an AI-enabled framework that accelerates MD simulation while maintaining high predictive accuracy. In order to extend AccelMD's applicability and flexibility, we also developed an extra preprocessing stage, which allows the framework to adapt to various protein inputs automatically. Compared to conventional MD, the AccelMD framework achieves 47.59 X speedup on average (max: 63.78 X, min: 30.61 X). Regarding the prediction accuracy of protein structures, the average Mean Absolute Error (MAE) and TMScore values of AccelMD are 0.0058 and 0.99980, respectively. Our empirical experiments exihibit the robustness of AccelMD framework for the study of complex, long-timescale molecular interactions.

The rise of Edge AI necessitates energy-efficient models like Spiking Neural Networks (SNNs), often trained using Federated Learning (FL) to preserve data privacy. However, FL is vulnerable to Byzantine attacks, where malicious clients disrupt training. While SNNs offer potential energy benefits due to their event-driven nature, their unique training mechanisms, particularly the use of surrogate gradients to handle non-differentiable spike events, raise questions about their inherent robustness in adversarial FL settings. We evaluate the robustness of SNNs employing 5 surrogate gradients (distinct by function shape) against 7 diverse Byzantine attacks and assess recovery potential using 5 robust aggregation rules (AGRs). Our extensive experiments (1032 runs) reveal that SNNs are not universally more robust than ANNs; they show resilience to certain structured attacks (e.g., MinMax) but vulnerability to others (e.g., Label Flip). We find a moderate positive correlation between surrogate gradient choice and recovery effectiveness using AGRs, with Triangle and Rectangle surrogates often enabling better recovery, though this advantage is context-dependent. Our results underscore that robust AGRs (like DnC and RFA) are essential for mitigating attacks in SNN-based FL, regardless of the surrogate gradient used. We conclude that achieving reliable SNN deployment in adversarial FL requires a holistic, context-aware approach, carefully considering the interplay between network type, surrogate gradient, threat model, and defense mechanisms. Our code is open-sourced for reproducibility1.

Spiking Neural Networks (SNN), which offer exceptional energy efficiency for inference, and Federated Learning (FL), which offers privacy-preserving training, is a rising area of interest that highly beneficial towards Internet of Things (IoT) devices. Despite this, research that tackles Byzantine attacks and bandwidth limitation in FL-SNN, both poses significant threats on model convergence and training times, still remains largely unexplored. In this paper, we first systematically evaluate the robustness of ANN and SNN in the FL context under four model-poisoning Byzantine attacks. We find that FL-SNN demonstrate better reliability than FL-ANN against most Byzantine attacks except MinMax. We then propose the$Top-\kappa$sparsification approach for better robustness in FL-SNN and to reduce communication overhead. Using this simple compression method, we observe ~40% accuracy enhancement in FL-SNN training under the lethal MinMax attack, leading to FL-SNN being more robust than FL-ANN in all four model-poisoning Byzantine attacks. This study highlights the dual benefits of FL-SNN with$Top-\kappa$sparsification in significantly reduce energy consumption and provide better robustness to Byzantine attacks for edge AI applications.

When multimodal AI systems increasingly utilize diverse data sources to achieve advanced understanding and interaction, they inevitably collect vast amounts of sensitive information, thus highlighting the urgent need for robust privacy safeguards, especially as these technologies expand into fields like healthcare, finance, and education. Existing research on data privacy in AI, encompassing adversarial training-based models, differential privacy-based models, and differentially private transform-based models, often neglects the inter-correlation inherent in multi-sensor data. To address this gap, we propose the differentiAl Private OnLine muLti-sensor data predictiOn model (APOLLO), which simultaneously considers intra-correlation and inter-correlation to enhance privacy protection while maintaining predictive performance. Under the proposed APOLLO frame-work, we design two implementations: APOLLO I, which ensures$\epsilon$-differential privacy by adding Laplace noise to each correlated data segment, and APOLLO II, which applies additional noise to make the concatenated multi-sensor data realize$\epsilon{-}$differential privacy. Furthermore, we conduct the theoretical analysis to reveal the relationship between performance influence and the privacy budget, providing guidelines for noise addition with the aim of achieving certified performance. Comprehensive experiments validate the effectiveness of the APOLLO model, establishing a new standard for privacy-preserving multi-sensor data prediction.

Deploying local AI models, such as Large Language Models (LLMs), to edge devices can substantially enhance devices’ independent capabilities, alleviate the server’s burden, and lower the response time. Owing to these tremendous potentials, many big tech companies have been actively promoting edge LLM evolution and released several lightweight Small Language Models (SLMs) to bridge this gap. However, SLMs currently only work well on limited real-world applications. We still have huge motivations to deploy more powerful (larger-scale) AI models on edge devices and enhance their smartness level. Unlike the conventional approaches for AI model compression, we investigate from activation sparsity. The activation sparsity method is orthogonal and combinable with existing techniques to maximize compression rate while maintaining great accuracy. According to statistics of open-source LLMs, their Feed-Forward Network (FFN) components typically comprise a large proportion of parameters (around $\tfrac{2}{3}$). This internal feature ensures that our FFN optimizations would have a better chance of achieving effective compression. Moreover, our findings are beneficial to general LLMs and are not restricted to ReLU-based models.This work systematically investigates the tradeoff between enforcing activation sparsity and perplexity (accuracy) on state-of-the-art LLMs. Our empirical analysis demonstrates that we can obtain around 50% of main memory and computing reductions for critical FFN components with negligible accuracy degradation. This extra 50% sparsity does not naturally exist in the current LLMs, which require tuning LLMs’ activation outputs by injecting zero-enforcing thresholds. To obtain the benefits of activation sparsity, we provide a guideline for the system architect for LLM prediction and prefetching. Moreover, we further verified the predictability of activation patterns in recent LLMs. The success prediction allows the system to prefetch the necessary weights while omitting the inactive ones and their successors (compress models from the memory’s perspective), therefore lowering cache/memory pollution and reducing LLM execution time on resource-constraint edge devices.

Spiking Neural Networks (SNNs) have recently gained significant interest in on-chip learning in embedded devices and emerged as an energy-efficient alternative to conventional Artificial Neural Networks (ANNs). However, to extend SNNs to a Federated Learning (FL) setting involving collaborative model training, the communication between the local devices and the remote server remains the bottleneck, which is often restricted and costly. In this paper, we first explore the inherent robustness of SNNs under noisy communication in FL. Building upon this foundation, we propose a novel Federated Learning with Top-κ Sparsification (FLTS) algorithm to reduce the bandwidth usage for FL training. We discover that the proposed scheme with SNNs allows more bandwidth savings compared to ANNs without impacting the model’s accuracy. Additionally, the number of parameters to be communicated can be reduced to as low as 6% of the size of the original model. We further improve the communication efficiency by enabling dynamic parameter compression during model training. Extensive experiment results demonstrate that our proposed algorithms significantly outperform the baselines in terms of communication cost and model accuracy and are promising for practical network-efficient FL with SNNs.

In teleradiology, medical images are transmitted to offsite radiologist for interpretation and the dictation report is sent back to the original site to aid timely diagnosis and proper patient care. Although teleradiology offers great benefits including time and cost efficiency, after-hour coverages, and staffing shortage management, there are some technical and operational limitations to overcome in reaching its full potential. We analyzed the current teleradiology workflow to identify inefficiencies. Image unavailability and delayed critical result communication stemmed from lack of system integration between teleradiology practice and healthcare institutions are among the most substantial factors causing prolonged turnaround time. In this paper, we propose a blockchain-based medical image sharing and automated critical-results notification platform to address the current limitation. We believe the proposed platform will enhance efficiency in workflow by eliminating the need for intermediaries and will benefit patients by eliminating the need for storing medical images in hard copies. While considerable progress was achieved, further research on governance and HIPAA compliance is required to optimize the adoption of the new application. Towards an idea to a working paradigm, we will implement the prototype during the next phase of our study.

Many tech stars like Netflix, Amazon, PayPal, eBay, and Twitter are evolving from monolithic to a microservice (containerization) architecture due to the benefits for Agile and DevOps teams. Microservices architecture can be applied to multiple industries, like IoT, using containerization. Since the IoT industry has exponential growth, universities' responsibility is to teach IoT with hands-on labs to minimize the gap between what the students learn and what is on-demand in the job market. There are many approaches in the containerization field, but they can be challenging to use without depth knowledge in virtualization and code encapsulation. After a deep analysis of the containerization challenges, in this paper, we present a cyberinfrastructure based on containers to solve the virtualization and code-encapsulation problems. The cyberinfrastructure will provide the necessary tools for data collection and code development and testing using an IoT Cluster. It is a web-based platform that allows users to securely go into containerization without spending time learning virtualization. Results show that our proposed cyberinfrastructure allows the creation and deployment of microservices in multiple IoT devices and ensures easy data collection for posterior cyberanalysis.

The convergence of the Internet of Things (IoT) and data analytics has great potential to accelerate knowledge discovery, while the traditional approach of centralized data collection then processing is becoming infeasible in many applications due to efficiency and privacy concerns. Federated learning (FL) has emerged as a new paradigm that enables model learning across distributed IoT devices without sharing raw data. However, previous works on FL are either relying on a single central server or fully decentralized. In this article, we propose a semihierarchical federated analytics framework combining the advantages of the above architectures. The proposed framework leverages multiple edge servers for aggregating updates from IoT devices and fusing learned model weights without the need of cloud or a central server. Besides, we develop a new local client update rule to further improve the communication efficiency by reducing the communication rounds between IoT devices and edge servers. We analyze the convergence properties of the presenting approach and investigate its characteristics considering the effects of varying parameters, unreliable links, and packet loss. The experimental results demonstrate the effectiveness of our proposed methodology in providing communication-efficient, robust, and fault-tolerant data analytics to IoT networks.

Edge computing is emerging as a new infrastructure for Internet-of-Things (IoT) networks by placing computation and analytics near to where data are generated. This article presents a novel data analytics framework for edge computing. The framework is based on a new decentralized algorithm, which enables all the nodes to obtain the global optimal model without sharing raw data. The resulting scheme executes in a hybrid mode: local IoT nodes send computed information to edge nodes. The edge nodes cooperate with each other by exchanging analytics with their neighbors only. The presenting approach is analyzed and evaluated on various applications and the experimental results demonstrate the effectiveness of the proposed methodology in providing fast data analytics to edge computing infrastructure.

We consider the decentralized consensus optimization problem arising from in-situ seismic tomography in large-scale sensor networks. Unlike traditional seismic imaging performed in a centralized location, each node in this setting privately holds an objective function and partial data. The goal of each node is to obtain the optimal solution of the whole seismic image, while communicating only with its immediate neighbors. We present a fast decentralized gradient descent method and prove that this new method can reach optimal convergence rate of O(1/k2) where k is the number of communication/iteration rounds. Extensive numerical experiments on synthetic and real-world sensor network seismic data demonstrate that the proposed algorithms significantly outperform existing methods.

This paper presents a new multi-objective microgrid reconfiguration problem formulation. Unlike existing distribution system or microgrid reconfiguration algorithms, we consider the effect of uncertainty arising from the renewable energy generation and investigate the tradeoff between the invented index measuring the reliability of reconfiguration and the total load served. The resulting optimization problem is computationally prohibitive due to the binary circuit breaker variables and the probability constraint accounting for the uncertainty of renewable generation. Nevertheless, a semidefinite programming (SDP) reformulation is developed based on convex relaxation techniques and the scenario-based approximation. Furthermore, weighted-sum method is applied in the reformulation and we eventually obtain the Pareto solution points of the microgrid reconfiguration. Numerical tests validate the intrinsic tradeoff between the two objectives and demonstrate the effectiveness of the proposed solution methodology.