Search

Feature interaction is crucial in predictive machine learning models, as it captures the relationships between features that influence model performance. In this work, we focus on pairwise interactions and investigate their importance in constructing feature graphs for Graph Neural Networks (GNNs). Rather than proposing new methods, we leverage existing GNN models and tools to explore the relationship between feature graph structures and their effectiveness in modeling interactions. Through experiments on synthesized datasets, we uncover that edges between interacting features are important for enabling GNNs to model feature interactions effectively. We also observe that including non-interaction edges can act as noise, degrading model performance. Furthermore, we provide theoretical support for sparse feature graph selection using the Minimum Description Length (MDL) principle. We prove that feature graphs retaining only necessary interaction edges yield a more efficient and interpretable representation than complete graphs, aligning with Occam's Razor. Our findings offer both theoretical insights and practical guidelines for designing feature graphs that improve the performance and interpretability of GNN models., Comment: This is the draft before submitting to any journal

This paper presents a wind farm layout optimization framework that integrates polynomial chaos expansion, a Kriging model, and the expected improvement algorithm. The proposed framework addresses the computational challenges associated with high-fidelity wind farm simulations by significantly reducing the number of function evaluations required for accurate annual energy production predictions. The polynomial chaos expansion-based prediction method achieves exceptional accuracy with reduced computational cost for over 96%, significantly lowering the expense of training the ensuing surrogate model. The Kriging model, combined with a genetic algorithm, is used for surrogate-based optimization, achieving comparable performance to direct optimization at a much-reduced computational cost. The integration of the expected improvement algorithm enhances the global optimization capability of the framework, allowing it to escape local optima and achieve results that are either nearly identical to or even outperform those obtained through direct optimization. The feasibility of the polynomial chaos expansion-Kriging framework is demonstrated through four case studies, including the optimization of wind farms with 8, 16, and 32 turbines using low-fidelity wake models, and a high-fidelity case using computational fluid dynamics simulations. The results show that the proposed framework is highly effective in optimizing wind farm layouts, significantly reducing computational costs while maintaining or improving the accuracy of annual energy production predictions.

Federated Foundation Models (FedFMs) represent a distributed learning paradigm that fuses general competences of foundation models as well as privacy-preserving capabilities of federated learning. This combination allows the large foundation models and the small local domain models at the remote clients to learn from each other in a teacher-student learning setting. This paper provides a comprehensive summary of the ten challenging problems inherent in FedFMs, encompassing foundational theory, utilization of private data, continual learning, unlearning, Non-IID and graph data, bidirectional knowledge transfer, incentive mechanism design, game mechanism design, model watermarking, and efficiency. The ten challenging problems manifest in five pivotal aspects: ``Foundational Theory," which aims to establish a coherent and unifying theoretical framework for FedFMs. ``Data," addressing the difficulties in leveraging domain-specific knowledge from private data while maintaining privacy; ``Heterogeneity," examining variations in data, model, and computational resources across clients; ``Security and Privacy," focusing on defenses against malicious attacks and model theft; and ``Efficiency," highlighting the need for improvements in training, communication, and parameter efficiency. For each problem, we offer a clear mathematical definition on the objective function, analyze existing methods, and discuss the key challenges and potential solutions. This in-depth exploration aims to advance the theoretical foundations of FedFMs, guide practical implementations, and inspire future research to overcome these obstacles, thereby enabling the robust, efficient, and privacy-preserving FedFMs in various real-world applications.

The second Hot QCD Matter 2024 conference at IIT Mandi focused on various ongoing topics in high-energy heavy-ion collisions, encompassing theoretical and experimental perspectives. This proceedings volume includes 19 contributions that collectively explore diverse aspects of the bulk properties of hot QCD matter. The topics encompass the dynamics of electromagnetic fields, transport properties, hadronic matter, spin hydrodynamics, and the role of conserved charges in high-energy environments. These studies significantly enhance our understanding of the complex dynamics of hot QCD matter, the quark-gluon plasma (QGP) formed in high-energy nuclear collisions. Advances in theoretical frameworks, including hydrodynamics, spin dynamics, and fluctuation studies, aim to improve theoretical calculations and refine our knowledge of the thermodynamic properties of strongly interacting matter. Experimental efforts, such as those conducted by the ALICE and STAR collaborations, play a vital role in validating these theoretical predictions and deepening our insight into the QCD phase diagram, collectivity in small systems, and the early-stage behavior of strongly interacting matter. Combining theoretical models with experimental observations offers a comprehensive understanding of the extreme conditions encountered in relativistic heavy-ion and proton-proton collisions., Comment: Compilation of the 19 contributions in Bulk Matter presented at the second 'Hot QCD Matter 2024 Conference' held from July 1-3, 2024, organized by IIT Mandi, Himachal Pradesh, India

This paper introduces SmartBSP, an advanced self-supervised learning framework for real-time path planning and obstacle avoidance in autonomous robotics navigating through complex environments. The proposed system integrates Proximal Policy Optimization (PPO) with Convolutional Neural Networks (CNN) and Actor-Critic architecture to process limited LIDAR inputs and compute spatial decision-making probabilities. The robot's perceptual field is discretized into a grid format, which the CNN analyzes to produce a spatial probability distribution. During the training process a nuanced cost function is minimized that accounts for path curvature, endpoint proximity, and obstacle avoidance. Simulations results in different scenarios validate the algorithm's resilience and adaptability across diverse operational scenarios. Subsequently, Real-time experiments, employing the Robot Operating System (ROS), were carried out to assess the efficacy of the proposed algorithm.

The Harris criterion imposes a constraint on the critical behavior of a system upon introduction of new disorder, based on its dimension $d$ and localization length exponent $\nu$. It states that the new disorder can be relevant only if $d \nu < 2$. We analyze the applicability of the Harris criterion to the network disorder formulated in the paper [I. A. Gruzberg, A. Kl\"umper, W. Nuding, and A. Sedrakyan, Phys. Rev. B 95, 125414 (2017)] and show that the fluctuations of the geometry are relevant despite $d \nu> 2$, implying that Harris criterion should be modified. We have observed that the fluctuations of the critical point in different quenched configurations of disordered network blocks is of order $L^0$, i.e.~it does not depend on block size $L$ in contrast to the expectation based on the Harris criterion that they should decrease as $L^{-d/2}$ according to the central limit theorem. Since $L^0 > (x-x_c)$ is always satisfied near the critical point, the mentioned network disorder is relevant and the critical indices of the system can be changed. We have also shown that our disordered network is fundamentally different from Voronoi-Delaunay and dynamically triangulated random lattices: the probability of higher connectivity in our network decreases in a power law as opposed to an exponential, indicating that we are dealing with a ``scale free" network, such as the Internet, protein-protein interactions, etc.

To understand the impact of vaccination, we consider a SIVR (susceptible-infected-vaccinated-recovered) model which combines impulsive vaccination into the classical SIR model. The final size is firstly defined and estimated, and then the peak value and peak time are considered. Finally, the critical times for a given infected number is studied, and it can be used to define and estimate the stopping time. Our results extend those for the well-understood SIR model., Comment: 16 pages, 5 figures

Current methods for microplastic identification in water samples are costly and require expert analysis. Here, we propose a deep learning segmentation model to automatically identify microplastics in microscopic images. We labeled images of microplastic from the Moore Institute for Plastic Pollution Research and employ a Generative Adversarial Network (GAN) to supplement and generate diverse training data. To verify the validity of the generated data, we conducted a reader study where an expert was able to discern the generated microplastic from real microplastic at a rate of 68 percent. Our segmentation model trained on the combined data achieved an F1-Score of 0.91 on a diverse dataset, compared to the model without generated data's 0.82. With our findings we aim to enhance the ability of both experts and citizens to detect microplastic across diverse ecological contexts, thereby improving the cost and accessibility of microplastic analysis., Comment: 6 pages one figure

Off-axis digital holographic microscopy is a high-throughput, label-free imaging technology that provides three-dimensional, high-resolution information about samples, particularly useful in large-scale cellular imaging. However, the hologram reconstruction process poses a significant bottleneck for timely data analysis. To address this challenge, we propose a novel reconstruction approach that integrates deep learning with the physical principles of off-axis holography. We initialized part of the network weights based on the physical principle and then fine-tuned them via weakly supersized learning. Our off-axis hologram network (OAH-Net) retrieves phase and amplitude images with errors that fall within the measurement error range attributable to hardware, and its reconstruction speed significantly surpasses the microscope's acquisition rate. Crucially, OAH-Net demonstrates remarkable external generalization capabilities on unseen samples with distinct patterns and can be seamlessly integrated with other models for downstream tasks to achieve end-to-end real-time hologram analysis. This capability further expands off-axis holography's applications in both biological and medical studies., Comment: 11 pages, 4 figures

This paper will explore the design and implementation of quantum networks in space integrated with quantum networks on Earth. We propose a three-layer approach, involving GEO and LEO satellites integrated with terrestrial ground stations. We first analyze the channel conditions between the three layers, and then highlight the key role of LEO satellites in the integrated space-terrestrial system - namely the source of entanglement distribution between specified terrestrial stations via direct downlink quantum-optical channels. The GEO satellites in the considered system are used primarily as coordination stations, managing and directing the LEO satellites regarding the positioning and timing of entanglement distribution. Complexity, in the form of entanglement distillation and quantum-state correction, is concentrated at the terrestrial stations, and teleportation is used as the primary quantum channel in the LEO uplinks and the inter-terrestrial channels. Although our designs are futuristic in that they assume limited quantum memory at the transceivers, we also discuss some near-term uses of our network in which no quantum memory is available., Comment: Published: IEEE Communications Magazine

Low-Latency and Low-Power Edge AI is essential for Virtual Reality and Augmented Reality applications. Recent advances show that hybrid models, combining convolution layers (CNN) and transformers (ViT), often achieve superior accuracy/performance tradeoff on various computer vision and machine learning (ML) tasks. However, hybrid ML models can pose system challenges for latency and energy-efficiency due to their diverse nature in dataflow and memory access patterns. In this work, we leverage the architecture heterogeneity from Neural Processing Units (NPU) and Compute-In-Memory (CIM) and perform diverse execution schemas to efficiently execute these hybrid models. We also introduce H4H-NAS, a Neural Architecture Search framework to design efficient hybrid CNN/ViT models for heterogeneous edge systems with both NPU and CIM. Our H4H-NAS approach is powered by a performance estimator built with NPU performance results measured on real silicon, and CIM performance based on industry IPs. H4H-NAS searches hybrid CNN/ViT models with fine granularity and achieves significant (up to 1.34%) top-1 accuracy improvement on ImageNet dataset. Moreover, results from our Algo/HW co-design reveal up to 56.08% overall latency and 41.72% energy improvements by introducing such heterogeneous computing over baseline solutions. The framework guides the design of hybrid network architectures and system architectures of NPU+CIM heterogeneous systems.

Systematic and comparative research on Lorenz ratios for graphene and nonrelativistic systems has been studied to identify their Wiedemann-Franz law violation domain. Fermi energy and temperature are the main governing parameters for deciding the values of the Lorenz ratio, which is basically thermal conductivity divided by electrical conductivity times temperature times Lorenz number. Metals as three-dimensional nonrelativistic electron gas locate at higher Fermi-energy by temperature domain, where Lorenz ratio remains one. Hence, they obey the Wiedemann-Franz law. By creating higher doping in a two-dimensional graphene system, one can again reach a higher Fermi-energy by temperature domain and get a constant Lorenz ratio. For both graphene and non-relativistic systems, the Lorenz ratio goes below one if we go lower Fermi-energy by temperature domain, which is possible for the graphene system by decreasing the doping concentration. Experimentally observed greater than one Lorenz ratio in this lower Fermi-energy by temperature domain or Dirac Fluid domain indicates that non-fluid expressions of Lorenz ratio should be replaced by fluid-type expressions. We have noticed a divergent trend of Lorenz ratio in the Dirac Fluid domain using its fluid-type expression, and it matches with the trend of experimental data., Comment: 14 pages, 4 figures

WalledEval is a comprehensive AI safety testing toolkit designed to evaluate large language models (LLMs). It accommodates a diverse range of models, including both open-weight and API-based ones, and features over 35 safety benchmarks covering areas such as multilingual safety, exaggerated safety, and prompt injections. The framework supports both LLM and judge benchmarking and incorporates custom mutators to test safety against various text-style mutations, such as future tense and paraphrasing. Additionally, WalledEval introduces WalledGuard, a new, small, and performant content moderation tool, and two datasets: SGXSTest and HIXSTest, which serve as benchmarks for assessing the exaggerated safety of LLMs and judges in cultural contexts. We make WalledEval publicly available at https://github.com/walledai/walledeval., Comment: Under review

In this paper we propose a new $S$-matrix approach to numerical simulations of network models and apply it to random networks that we proposed in a previous work 10.1103/PhysRevB.95.125414. Random networks are modifications of the Chalker-Coddington (CC) model for the integer quantum Hall transition that more faithfully capture the physics of electrons moving in a strong magnetic field and a smooth disorder potential. The new method has considerable advantages compared to the transfer matrix approach, and gives the value $\nu \approx 2.4$ for the critical exponent of the localization length in a random network. This finding confirms our previous result and is surprisingly close to the experimental value $\nu_{\text{exp}} \approx 2.38$ observed at the integer quantum Hall transition but substantially different from the CC value $\nu_\text{CC} \approx 2.6$.

The advent of Federated Learning (FL) highlights the practical necessity for the right to be forgotten for all clients, allowing them to request data deletion from the machine learning models service provider. This necessity has spurred a growing demand for Federated Unlearning (FU). Feature unlearning has gained considerable attention due to its applications in unlearning sensitive, backdoor, and biased features. Existing methods employ the influence function to achieve feature unlearning, which is impractical for FL as it necessitates the participation of other clients, if not all, in the unlearning process. Furthermore, current research lacks an evaluation of the effectiveness of feature unlearning. To address these limitations, we define feature sensitivity in evaluating feature unlearning according to Lipschitz continuity. This metric characterizes the model outputs rate of change or sensitivity to perturbations in the input feature. We then propose an effective federated feature unlearning framework called Ferrari, which minimizes feature sensitivity. Extensive experimental results and theoretical analysis demonstrate the effectiveness of Ferrari across various feature unlearning scenarios, including sensitive, backdoor, and biased features. The code is publicly available at https://github.com/OngWinKent/Federated-Feature-Unlearning, Comment: TLDR: The need for a "right to be forgotten" in Federated Learning has led to the development of the Ferrari framework, which efficiently unlearns sensitive features using a Lipschitz continuity-based metric, proven effective in extensive testing. Accepted at NeurIPS 2024

Current vision-language pre-training (VLP) methodologies predominantly depend on paired image-text datasets, a resource that is challenging to acquire in radiology due to privacy considerations and labelling complexities. Data augmentation provides a practical solution to overcome the issue of data scarcity, however, most augmentation methods exhibit a limited focus, prioritising either image or text augmentation exclusively. Acknowledging this limitation, our objective is to devise a framework capable of concurrently augmenting medical image and text data. We design a Pairwise Augmentation (PairAug) approach that contains an Inter-patient Augmentation (InterAug) branch and an Intra-patient Augmentation (IntraAug) branch. Specifically, the InterAug branch of our approach generates radiology images using synthesised yet plausible reports derived from a Large Language Model (LLM). The generated pairs can be considered a collection of new patient cases since they are artificially created and may not exist in the original dataset. In contrast, the IntraAug branch uses newly generated reports to manipulate images. This process allows us to create new paired data for each individual with diverse medical conditions. Our extensive experiments on various downstream tasks covering medical image classification zero-shot and fine-tuning analysis demonstrate that our PairAug, concurrently expanding both image and text data, substantially outperforms image-/text-only expansion baselines and advanced medical VLP baselines. Our code is released at \url{https://github.com/YtongXie/PairAug}., Comment: Accepted to CVPR2024

Multispectral transmission imaging provides strong benefits for early breast cancer screening. The frame accumulation method addresses the challenge of low grayscale and signal-to-noise ratio resulting from the strong absorption and scattering of light by breast tissue. This method introduces redundancy in data while improving the grayscale and signal-to-noise ratio of the image. Existing terraced compression algorithms effectively eliminate the data redundancy introduced by frame accumulation but necessitate significant time for manual debugging of threshold values. Hence, this paper proposes an improved terrace compression algorithm. The algorithm necessitates solely the input of the desired heterogeneous body size and autonomously calculates the optimal area threshold and gradient threshold by counting the grayscale and combining its distribution. Experimental acquisition involved multi-wavelength images of heterogeneous bodies exhibiting diverse textures, depths, and thicknesses. Subsequently, the method was applied after pre-processing to determine the thresholds for terraced compression at each wavelength, coupled with a window function for multi-dimensional image clustering. The results illustrate the method's efficacy in detecting and identifying various heterogeneous body types, depths, and thicknesses. This approach is expected to accurately identify the locations and types of breast tumors in the future, thus providing a more dependable tool for early breast cancer screening.