Search

Three-dimensional (3D) spot beamfocusing (SBF), in contrast to conventional angular-domain beamforming, concentrates radiating power within a very small volume in both radial and angular domains in the near-field zone. Recently the implementation of channel-state-information (CSI)-independent machine learning (ML)-based approaches have been developed for effective SBF using extremely large-scale programmable metasurface (ELPMs). These methods involve dividing the ELPMs into subarrays and independently training them with Deep Reinforcement Learning to jointly focus the beam at the desired focal point (DFP). This paper explores near-field SBF using ELPMs, addressing challenges associated with lengthy training times resulting from independent training of subarrays. To achieve a faster CSI-independent solution, inspired by the correlation between the beamfocusing matrices of the subarrays, we leverage transfer learning techniques. First, we introduce a novel similarity criterion based on the phase distribution image (PDI) of subarray apertures. Then we devise a subarray policy propagation scheme that transfers the knowledge from trained to untrained subarrays. We further enhance learning by introducing quasi-liquid layers as a revised version of the adaptive policy reuse technique. We show through simulations that the proposed scheme improves the training speed about 5 times. Furthermore, for dynamic DFP management, we devised a DFP policy blending process, which augments the convergence rate up to 8-fold.

In this brief, we present an enhanced privacy-preserving distributed estimation algorithm, referred to as the ``Double-Private Algorithm," which combines the principles of both differential privacy (DP) and cryptography. The proposed algorithm enhances privacy by introducing DP noise into the intermediate estimations of neighboring nodes. Additionally, we employ an inverse of a closed-form reproducible proportionate gain matrix as the cryptographic key matrix to fortify the privacy protection within the proposed double private algorithm. \textcolor{blue}{We improve the algorithm by transmitting alternative variable vectors instead of raw measurements, resulting in enhanced key matrix reconstruction performance. This innovative approach mitigate noise impact, enhancing overall algorithm effectiveness.} We also establish an upper bound for the norm of the error between the non-private Diffusion Least Mean Square (DLMS) algorithm and our double private algorithm. Further, we determine a sufficient condition for the step-size to ensure the mean convergence of the proposed algorithm. Simulation results demonstrate the effectiveness of the proposed algorithm, particularly its ability to attain the final Mean Square Deviation (MSD) comparable to that of the non-private DLMS.

Branched actin networks exert pushing forces in eukaryotic cells, and adapt their stiffness to their environment. The physical basis for their mechanics and adaptability is however not understood. Indeed, here we show that their high density and low connectivity place them outside the scope of standard elastic network models for actin. We combine high-precision mechanical experiments, molecular dynamics simulations and a mean-field elastic theory to show that they are instead dominated by the proliferation of interfilament contacts under compression. This places branched actin in the same category as undercoordinated, fibrous materials such as sheep's wool. When the network is grown under force, filaments entangle as if knitted together and trap contacts in their structure. Trapped contacts play a similar role as crosslinkers in rigidifying the network, and are thus key to its active adaptive mechanics., Comment: 7 pages, 5 figures

The merging black hole binaries detected by the LIGO-Virgo-KAGRA (LVK) gravitational-wave observatories, may help us shed light on how such binaries form. In addition, these detections can help us probe the hypothesized primordial black holes, a candidate for the observed abundance of dark matter. In this work, we study the black-hole mass distribution obtained from the LVK binary black hole merger events. We obtain that distribution by first associating a skewed normal distribution to each event detected with a signal to noise ratio (SNR) $>$ 8 and then summing all such distributions. We also simulate black hole binaries from two separate populations of merging binaries. One of these is a stellar-origin population that follows a mass-distribution similar to the zero-age mass function of stars. The second population of black holes follows a Gaussian mass-distribution. Such a distribution could approximate a population of black hole binaries formed from earlier black hole mergers in dense stellar environments, or binaries of primordial black holes. For those populations, we evaluate the number of detectable events and fit their combination to the LVK observations. In our work, we rely on a wide range of stellar-origin black-hole mass distributions. We find that the observed LVK events can be fitted much better by the combination of such a stellar-origin mass distribution and a Gaussian distribution, than by the stellar-origin mass distribution alone., Comment: 14 pages, 11 figures, 5 tables

We study slow-roll inflation as a common feature in Metric-affine Theories of Gravity. In particular, we take into account extended metric, teleparallel and symmetric-teleparallel theories of gravity, based on different geometric invariants, discussing analogies and differences. The analysis for each model is performed in two approaches. First, we focus on the {\it potential-slow-roll approach} by studying the reconstructed potentials for different forms of the extended models related to the considered gravitational theory in the Einstein frame. Secondly, we investigate the {\it Hubble-slow-roll approach} for some conventional inflationary potentials related to the specific extended model in the Jordan frame. We compare all results with cosmic microwave background anisotropy observations coming from Planck 2018 and BICEP2/Keck array satellites in order to find the observational constraints on the parameters space of the models as well as their prediction from the spectral parameters. Eventually, we attempt to present a qualitative comparison between three classes of considered modified gravities using the obtained inflationary results. The aim is to select, in principle, cosmological signatures capable of discriminating among concurrent models in view to point out the representation of gravity, according with geometric invariants, which better addresses the early universe dynamics., Comment: 56 pages, 23 figures, 2 tables

Benchmarks are essential in the design of modern HPC installations, as they define key aspects of system components. Beyond synthetic workloads, it is crucial to include real applications that represent user requirements into benchmark suites, to guarantee high usability and widespread adoption of a new system. Given the significant investments in leadership-class supercomputers of the exascale era, this is even more important and necessitates alignment with a vision of Open Science and reproducibility. In this work, we present the JUPITER Benchmark Suite, which incorporates 16 applications from various domains. It was designed for and used in the procurement of JUPITER, the first European exascale supercomputer. We identify requirements and challenges and outline the project and software infrastructure setup. We provide descriptions and scalability studies of selected applications and a set of key takeaways. The JUPITER Benchmark Suite is released as open source software with this work at https://github.com/FZJ-JSC/jubench., Comment: To be published in Proceedings of The International Conference for High Performance Computing Networking, Storage, and Analysis (SC '24) (2024)

We develop a new approach for efficient and scalable simulations of measurement and control of quantum systems built upon existing phase-space methods, namely the Truncated Wigner Approximation (TWA). We benchmark against existing particle-filter methods by simulating measurement based feedback cooling in a two-mode system, whose low-dimensional nature permits a computation of an exact solution. The advantage of our method is multi-mode scalability, which we demonstrate through the first successful simulation of measurement-based feedback cooling of an incoherent quasi-1D thermal ensemble to quantum degeneracy. As the underlying principle of our approach exploits a general correspondence between measurement and coherent feedback, we anticipate it is also applicable across a broad range of other quantum control scenarios., Comment: 11 pages, 4 figures. 03/09/24 Version: Minor edits in references, fixed author list formatting

Deep supervised learning models require high volume of labeled data to attain sufficiently good results. Although, the practice of gathering and annotating such big data is costly and laborious. Recently, the application of self supervised learning (SSL) in vision tasks has gained significant attention. The intuition behind SSL is to exploit the synchronous relationships within the data as a form of self-supervision, which can be versatile. In the current big data era, most of the data is unlabeled, and the success of SSL thus relies in finding ways to utilize this vast amount of unlabeled data available. Thus it is better for deep learning algorithms to reduce reliance on human supervision and instead focus on self-supervision based on the inherent relationships within the data. With the advent of ViTs, which have achieved remarkable results in computer vision, it is crucial to explore and understand the various SSL mechanisms employed for training these models specifically in scenarios where there is limited labelled data available. In this survey, we develop a comprehensive taxonomy of systematically classifying the SSL techniques based upon their representations and pre-training tasks being applied. Additionally, we discuss the motivations behind SSL, review popular pre-training tasks, and highlight the challenges and advancements in this field. Furthermore, we present a comparative analysis of different SSL methods, evaluate their strengths and limitations, and identify potential avenues for future research., Comment: 34 Pages, 5 Figures, 7 Tables

As is widely known, the computational speed and power consumption are two critical parameters in microprocessor design. A solution for these issues is the application specific instruction set processor (ASIP) methodology, which can improve speed and reduce power consumption of the general purpose processor (GPP) technique. In ASIP, changing the instruction set architecture (ISA) of the processor will lead to alter the number and the mean time of accesses to the cache memory. This issue has a direct impact on the processor energy consumption. In this work, we study the impacts of extended ISA on the energy consumption of the extended ISA processor. Also, we demonstrate the extended ISA let the designer to reduce the cache size in order to minimize the energy consumption while meeting performance constraint.

Measuring the growth of structure is a powerful probe for studying the dark sector, especially in light of the $\sigma_8$ tension between primary CMB anisotropy and low-redshift surveys. This paper provides a new measurement of the amplitude of the matter power spectrum, $\sigma_8$, using galaxy-galaxy and galaxy-CMB lensing power spectra of Dark Energy Spectroscopic Instrument Legacy Imaging Surveys Emission-Line Galaxies and the $\textit{Planck}$ 2018 CMB lensing map. We create an ELG catalog composed of $27$ million galaxies and with a purity of $85\%$, covering a redshift range $0 < z < 3$, with $z_{\rm mean} = 1.09$. We implement several novel systematic corrections, such as jointly modeling the contribution of imaging systematics and photometric redshift uncertainties to the covariance matrix. We also study the impacts of various dust maps on cosmological parameter inference. We measure the cross-power spectra over $f_{\rm sky} = 0.25$ with a signal-to-background ratio of up to $ 30\sigma$. We find that the choice of dust maps to account for imaging systematics in estimating the ELG overdensity field has a significant impact on the final estimated values of $\sigma_8$ and $\Omega_{\rm M}$, with far-infrared emission-based dust maps preferring $\sigma_8$ to be as low as $0.702 \pm 0.030$, and stellar-reddening-based dust maps preferring as high as $0.719 \pm 0.030$. The highest preferred value is at $\sim 3 \sigma$ tension with the $\textit{Planck}$ primary anisotropy results. These findings indicate a need for tomographic analyses at high redshifts and joint modeling of systematics., Comment: 50 pages, 24 figures (figure data can be obtained at https://zenodo.org/records/13381499)

In an era where digital text is proliferating at an unprecedented rate, efficient summarization tools are becoming indispensable. While Large Language Models (LLMs) have been successfully applied in various NLP tasks, their role in extractive text summarization remains underexplored. This paper introduces EYEGLAXS (Easy Yet Efficient larGe LAnguage model for eXtractive Summarization), a framework that leverages LLMs, specifically LLAMA2-7B and ChatGLM2-6B, for extractive summarization of lengthy text documents. Instead of abstractive methods, which often suffer from issues like factual inaccuracies and hallucinations, EYEGLAXS focuses on extractive summarization to ensure factual and grammatical integrity. Utilizing state-of-the-art techniques such as Flash Attention and Parameter-Efficient Fine-Tuning (PEFT), EYEGLAXS addresses the computational and resource challenges typically associated with LLMs. The system sets new performance benchmarks on well-known datasets like PubMed and ArXiv. Furthermore, we extend our research through additional analyses that explore the adaptability of LLMs in handling different sequence lengths and their efficiency in training on smaller datasets. These contributions not only set a new standard in the field but also open up promising avenues for future research in extractive text summarization.

The cell-free massive multiple-input multiple-output (CF-mMIMO) systems are crucial for 6G development due to their high spectral efficiency and uniform user-experienced data rates. A key aspect of CF-mMIMO is user association (UA) and optimal cluster formation. Traditional methods focusing solely on communication-related metrics fall short in this context, as sensing is becoming integral to 6G. This study delves into a framework for joint radar and communication (JRC) in CF-mMIMO systems and investigates JRC-based UA techniques. We propose a novel method to optimize UA, enhancing both communication spectral efficiency and sensing accuracy. Existing literature has not explored this dual requirement integration for UA. Our proposed two-step scheme optimizes UA clusters for both communication and sensing. The first step involves selecting access points (APs) based on channel quality, followed by a second step that further refines the selection by choosing APs from the initial group that are also optimal for sensing. We utilize the signal-clutter plus noise ratio to exclude APs with clutter in front of the user equipment (UE) and the AP view angle, ensuring that radar echoes are received only from the specific UE, not the surrounding clutter. Theoretical analysis and simulations demonstrate that the same APs optimized for communication are not necessarily optimal for sensing, highlighting the need for schemes that incorporate sensing requirements in UA. The results show the effectiveness of the proposed method, showing its potential to improve CF-mMIMO system performance in JRC scenarios.

Multimodal foundation models serve numerous applications at the intersection of vision and language. Still, despite being pretrained on extensive data, they become outdated over time. To keep models updated, research into continual pretraining mainly explores scenarios with either (1) infrequent, indiscriminate updates on large-scale new data, or (2) frequent, sample-level updates. However, practical model deployment often operates in the gap between these two limit cases, as real-world applications often demand adaptation to specific subdomains, tasks or concepts -- spread over the entire, varying life cycle of a model. In this work, we complement current perspectives on continual pretraining through a research test bed as well as provide comprehensive guidance for effective continual model updates in such scenarios. We first introduce FoMo-in-Flux, a continual multimodal pretraining benchmark with realistic compute constraints and practical deployment requirements, constructed over 63 datasets with diverse visual and semantic coverage. Using FoMo-in-Flux, we explore the complex landscape of practical continual pretraining through multiple perspectives: (1) A data-centric investigation of data mixtures and stream orderings that emulate real-world deployment situations, (2) a method-centric investigation ranging from simple fine-tuning and traditional continual learning strategies to parameter-efficient updates and model merging, (3) meta learning rate schedules and mechanistic design choices, and (4) the influence of model and compute scaling. Together, our insights provide a practitioner's guide to continual multimodal pretraining for real-world deployment. Our benchmark and code is here: https://github.com/ExplainableML/fomo_in_flux., Comment: Technical Report. 52 pages. Shorter version published at the NeurIPS 2024 Dataset & Benchmarks track

Despite decades of research on blood flow, an analogue of Navier-Stokes equations that accurately describe blood flow properties has not been established yet. The reason behind this is that the properties of blood flow seem \`a priori non universal as they depend on various factors such as global concentration of red blood cells (RBCs) and channel width. Here, we have discovered a universal law when the stress and strain rate are measured at a given local RBCs concentration. However, the local concentration must be determined in order to close the problem. We propose a non-local diffusion equation of RBCs concentration that agrees with the full simulation. The universal law is exemplified for both shear and pressure driven flows. While the theory is restricted to a simplistic geometry (straight channel) it provides a fundamental basis for future research on blood flow dynamics and could lead to the development of a new theory that accurately describes blood flow properties under various conditions, such as in complex vascular networks.

Federated Learning (FL) offers a promising approach for collaborative machine learning across distributed devices. However, its adoption is hindered by the complexity of building reliable communication architectures and the need for expertise in both machine learning and network programming. This paper presents a comprehensive solution that simplifies the orchestration of FL tasks while integrating intent-based automation. We develop a user-friendly web application supporting the federated averaging (FedAvg) algorithm, enabling users to configure parameters through an intuitive interface. The backend solution efficiently manages communication between the parameter server and edge nodes. We also implement model compression and scheduling algorithms to optimize FL performance. Furthermore, we explore intent-based automation in FL using a fine-tuned Language Model (LLM) trained on a tailored dataset, allowing users to conduct FL tasks using high-level prompts. We observe that the LLM-based automated solution achieves comparable test accuracy to the standard web-based solution while reducing transferred bytes by up to 64% and CPU time by up to 46% for FL tasks. Also, we leverage the neural architecture search (NAS) and hyperparameter optimization (HPO) using LLM to improve the performance. We observe that by using this approach test accuracy can be improved by 10-20% for the carried out FL tasks.

Predictive coding (PC) is an energy-based learning algorithm that performs iterative inference over network activities before updating weights. Recent work suggests that PC can converge in fewer learning steps than backpropagation thanks to its inference procedure. However, these advantages are not always observed, and the impact of PC inference on learning is not theoretically well understood. Here, we study the geometry of the PC energy landscape at the inference equilibrium of the network activities. For deep linear networks, we first show that the equilibrated energy is simply a rescaled mean squared error loss with a weight-dependent rescaling. We then prove that many highly degenerate (non-strict) saddles of the loss including the origin become much easier to escape (strict) in the equilibrated energy. Our theory is validated by experiments on both linear and non-linear networks. Based on these and other results, we conjecture that all the saddles of the equilibrated energy are strict. Overall, this work suggests that PC inference makes the loss landscape more benign and robust to vanishing gradients, while also highlighting the fundamental challenge of scaling PC to deeper models., Comment: 35 pages, 12 figures

Concentrating Solar Power (CSP) plants play a crucial role in the global transition towards sustainable energy. A key factor in ensuring the safe and efficient operation of CSP plants is the distribution of concentrated flux density on the receiver. However, the non-ideal flux density generated by individual heliostats can undermine the safety and efficiency of the power plant. The flux density from each heliostat is influenced by its precise surface profile, which includes factors such as canting and mirror errors. Accurately measuring these surface profiles for a large number of heliostats in operation is a formidable challenge. Consequently, control systems often rely on the assumption of ideal surface conditions, which compromises both safety and operational efficiency. In this study, we introduce inverse Deep Learning Ray Tracing (iDLR), an innovative method designed to predict heliostat surfaces based solely on target images obtained during heliostat calibration. Our simulation-based investigation demonstrates that sufficient information regarding the heliostat surface is retained in the flux density distribution of a single heliostat, enabling deep learning models to accurately predict the underlying surface with deflectometry-like precision for the majority of heliostats. Additionally, we assess the limitations of this method, particularly in relation to surface accuracy and resultant flux density predictions. Furthermore, we are presenting a new comprehensive heliostat model using Non-Uniform Rational B-Spline (NURBS) that has the potential to become the new State of the Art for heliostat surface parameterization. Our findings reveal that iDLR has significant potential to enhance CSP plant operations, potentially increasing the overall efficiency and energy output of the power plants.

Integrating robots into teams of humans is anticipated to bring significant capability improvements for tasks such as searching potentially hazardous buildings. Trust between humans and robots is recognized as a key enabler for human-robot teaming (HRT) activity: if trust during a mission falls below sufficient levels for cooperative tasks to be completed, it could critically affect success. Changes in trust could be particularly problematic in teams that have formed on an ad hoc basis (as might be expected in emergency situations) where team members may not have previously worked together. In such ad hoc teams, a foundational level of 'swift trust' may be fragile and challenging to sustain in the face of inevitable setbacks. We present results of an experiment focused on understanding trust building, violation and repair processes in ad hoc teams (one human and two robots). Trust violation occurred through robots becoming unresponsive, with limited communication and feedback. We perform exploratory analysis of a variety of data, including communications and performance logs, trust surveys and post-experiment interviews, toward understanding how autonomous systems can be designed into interdependent ad hoc human-robot teams where swift trust can be sustained.

Despite their widespread adoption, large language models (LLMs) remain prohibitive to use under resource constraints, with their ever growing sizes only increasing the barrier for use. One noted issue is the high latency associated with auto-regressive generation, rendering large LLMs use dependent on advanced computing infrastructure. Assisted decoding, where a smaller draft model guides a larger target model's generation, has helped alleviate this, but remains dependent on alignment between the two models. Thus if the draft model is insufficiently capable on some domain relative to the target model, performance can degrade. Alternatively, one can leverage multiple draft models to better cover the expertise of the target, but when multiple black-box draft models are available, selecting an assistant without details about its construction can be difficult. To better understand this decision making problem, we observe it as a contextual bandit, where a policy must choose a draft model based on a context. We show that even without prior knowledge of the draft models, creating an offline dataset from only outputs of independent draft/target models and training a policy over the alignment of these outputs can accelerate performance on multiple domains provided the candidates are effective. Further results show this to hold on various settings with multiple assisted decoding candidates, highlighting its flexibility and the advantageous role that such decision making can play., Comment: Published as a long paper at the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP). Official version of paper within conference proceedings is available at http://aclanthology.org/2024.emnlp-main.332

The emergence of synthetic data represents a pivotal shift in modern machine learning, offering a solution to satisfy the need for large volumes of data in domains where real data is scarce, highly private, or difficult to obtain. We investigate the feasibility of creating realistic, large-scale synthetic datasets of user-generated content, noting that such content is increasingly prevalent and a source of frequently sought information. Large language models (LLMs) offer a starting point for generating synthetic social media discussion threads, due to their ability to produce diverse responses that typify online interactions. However, as we demonstrate, straightforward application of LLMs yields limited success in capturing the complex structure of online discussions, and standard prompting mechanisms lack sufficient control. We therefore propose a multi-step generation process, predicated on the idea of creating compact representations of discussion threads, referred to as scaffolds. Our framework is generic yet adaptable to the unique characteristics of specific social media platforms. We demonstrate its feasibility using data from two distinct online discussion platforms. To address the fundamental challenge of ensuring the representativeness and realism of synthetic data, we propose a portfolio of evaluation measures to compare various instantiations of our framework.

Reward machines allow the definition of rewards for temporally extended tasks and behaviors. Specifying "informative" reward machines can be challenging. One way to address this is to generate reward machines from a high-level abstract description of the learning environment, using techniques such as AI planning. However, previous planning-based approaches generate a reward machine based on a single (sequential or partial-order) plan, and do not allow maximum flexibility to the learning agent. In this paper we propose a new approach to synthesising reward machines which is based on the set of partial order plans for a goal. We prove that learning using such "maximally permissive" reward machines results in higher rewards than learning using RMs based on a single plan. We present experimental results which support our theoretical claims by showing that our approach obtains higher rewards than the single-plan approach in practice., Comment: Paper accepted for publication at the European Conference on Artificial Intelligence (ECAI) 2024

Safe Reinforcement Learning (Safe RL) is one of the prevalently studied subcategories of trial-and-error-based methods with the intention to be deployed on real-world systems. In safe RL, the goal is to maximize reward performance while minimizing constraints, often achieved by setting bounds on constraint functions and utilizing the Lagrangian method. However, deploying Lagrangian-based safe RL in real-world scenarios is challenging due to the necessity of threshold fine-tuning, as imprecise adjustments may lead to suboptimal policy convergence. To mitigate this challenge, we propose a unified Lagrangian-based model-free architecture called Meta Soft Actor-Critic Lagrangian (Meta SAC-Lag). Meta SAC-Lag uses meta-gradient optimization to automatically update the safety-related hyperparameters. The proposed method is designed to address safe exploration and threshold adjustment with minimal hyperparameter tuning requirement. In our pipeline, the inner parameters are updated through the conventional formulation and the hyperparameters are adjusted using the meta-objectives which are defined based on the updated parameters. Our results show that the agent can reliably adjust the safety performance due to the relatively fast convergence rate of the safety threshold. We evaluate the performance of Meta SAC-Lag in five simulated environments against Lagrangian baselines, and the results demonstrate its capability to create synergy between parameters, yielding better or competitive results. Furthermore, we conduct a real-world experiment involving a robotic arm tasked with pouring coffee into a cup without spillage. Meta SAC-Lag is successfully trained to execute the task, while minimizing effort constraints., Comment: Main text accepted to the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2024, 10 pages, 4 figures, 3 tables

Data lakes are widely used to store extensive and heterogeneous datasets for advanced analytics. However, the unstructured nature of data in these repositories introduces complexities in exploiting them and extracting meaningful insights. This motivates the need of exploring efficient approaches for consolidating data lakes and deriving a common and unified schema. This paper introduces a practical data visualization and analysis approach rooted in Formal Concept Analysis (FCA) to systematically clean, organize, and design data structures within a data lake. We explore diverse data structures stored in our data lake at Infologic, including InfluxDB measurements and Elasticsearch indexes, aiming to derive conventions for a more accessible data model. Leveraging FCA, we represent data structures as objects, analyze the concept lattice, and present two strategies-top-down and bottom-up-to unify these structures and establish a common schema. Our methodology yields significant results, enabling the identification of common concepts in the data structures, such as resources along with their underlying shared fields (timestamp, type, usedRatio, etc.). Moreover, the number of distinct data structure field names is reduced by 54 percent (from 190 to 88) in the studied subset of our data lake. We achieve a complete coverage of 80 percent of data structures with only 34 distinct field names, a significant improvement from the initial 121 field names that were needed to reach such coverage. The paper provides insights into the Infologic ecosystem, problem formulation, exploration strategies, and presents both qualitative and quantitative results.

Implementing Machine Learning (ML) models on Field-Programmable Gate Arrays (FPGAs) is becoming increasingly popular across various domains as a low-latency and low-power solution that helps manage large data rates generated by continuously improving detectors. However, developing ML models for FPGAs is time-consuming, as optimization requires synthesis to evaluate FPGA area and latency, making the process slow and repetitive. This paper introduces a novel method to predict the resource utilization and inference latency of Neural Networks (NNs) before their synthesis and implementation on FPGA. We leverage HLS4ML, a tool-flow that helps translate NNs into high-level synthesis (HLS) code, to synthesize a diverse dataset of NN architectures and train resource utilization and inference latency predictors. While HLS4ML requires full synthesis to obtain resource and latency insights, our method uses trained regression models for immediate pre-synthesis predictions. The prediction models estimate the usage of Block RAM (BRAM), Digital Signal Processors (DSP), Flip-Flops (FF), and Look-Up Tables (LUT), as well as the inference clock cycles. The predictors were evaluated on both synthetic and existing benchmark architectures and demonstrated high accuracy with R2 scores ranging between 0.8 and 0.98 on the validation set and sMAPE values between 10% and 30%. Overall, our approach provides valuable preliminary insights, enabling users to quickly assess the feasibility and efficiency of NNs on FPGAs, accelerating the development and deployment processes. The open-source repository can be found at https://github.com/IMPETUS-UdeS/rule4ml, while the datasets are publicly available at https://borealisdata.ca/dataverse/rule4ml.

Classical probabilistic rounding error analysis is particularly well suited to stochastic rounding (SR), and it yields strong results when dealing with floating-point algorithms that rely heavily on summation. For many numerical linear algebra algorithms, one can prove probabilistic error bounds that grow as O(nu), where n is the problem size and u is the unit roundoff. These probabilistic bounds are asymptotically tighter than the worst-case ones, which grow as O(nu). For certain classes of algorithms, SR has been shown to be unbiased. However, all these results were derived under the assumption that SR is implemented exactly, which typically requires a number of random bits that is too large to be suitable for practical implementations. We investigate the effect of the number of random bits on the probabilistic rounding error analysis of SR. To this end, we introduce a new rounding mode, limited-precision SR. By taking into account the number r of random bits used, this new rounding mode matches hardware implementations accurately, unlike the ideal SR operator generally used in the literature. We show that this new rounding mode is biased and that the bias is a function of r. As r approaches infinity, however, the bias disappears, and limited-precision SR converges to the ideal, unbiased SR operator. We develop a novel model for probabilistic error analysis of algorithms employing SR. Several numerical examples corroborate our theoretical findings.

Minimizing negative impacts of Artificial Intelligent (AI) systems on human societies without human supervision requires them to be able to align with human values. However, most current work only addresses this issue from a technical point of view, e.g., improving current methods relying on reinforcement learning from human feedback, neglecting what it means and is required for alignment to occur. Here, we propose to distinguish strong and weak value alignment. Strong alignment requires cognitive abilities (either human-like or different from humans) such as understanding and reasoning about agents' intentions and their ability to causally produce desired effects. We argue that this is required for AI systems like large language models (LLMs) to be able to recognize situations presenting a risk that human values may be flouted. To illustrate this distinction, we present a series of prompts showing ChatGPT's, Gemini's and Copilot's failures to recognize some of these situations. We moreover analyze word embeddings to show that the nearest neighbors of some human values in LLMs differ from humans' semantic representations. We then propose a new thought experiment that we call "the Chinese room with a word transition dictionary", in extension of John Searle's famous proposal. We finally mention current promising research directions towards a weak alignment, which could produce statistically satisfying answers in a number of common situations, however so far without ensuring any truth value., Comment: Accepted for publication in Scientific Reports, special issue on AI aligment

In computer vision, the vision transformer (ViT) has increasingly superseded the convolutional neural network (CNN) for improved accuracy and robustness. However, ViT's large model sizes and high sample complexity make it difficult to train on resource-constrained edge devices. Split learning (SL) emerges as a viable solution, leveraging server-side resources to train ViTs while utilizing private data from distributed devices. However, SL requires additional information exchange for weight updates between the device and the server, which can be exposed to various attacks on private training data. To mitigate the risk of data breaches in classification tasks, inspired from the CutMix regularization, we propose a novel privacy-preserving SL framework that injects Gaussian noise into smashed data and mixes randomly chosen patches of smashed data across clients, coined DP-CutMixSL. Our analysis demonstrates that DP-CutMixSL is a differentially private (DP) mechanism that strengthens privacy protection against membership inference attacks during forward propagation. Through simulations, we show that DP-CutMixSL improves privacy protection against membership inference attacks, reconstruction attacks, and label inference attacks, while also improving accuracy compared to DP-SL and DP-MixSL., Comment: 23 pages, 11 figures, 8 tables, to be published in Transactions on Machine Learning Research (TMLR)

Artificial Intelligence (AI) advancements have enabled the development of Large Language Models (LLMs) that can perform a variety of tasks with remarkable semantic understanding and accuracy. ChatGPT is one such LLM that has gained significant attention due to its impressive capabilities for assisting in various knowledge-intensive tasks. Due to the knowledge-intensive nature of engineering secure software, ChatGPT's assistance is expected to be explored for security-related tasks during the development/evolution of software. To gain an understanding of the potential of ChatGPT as an emerging technology for supporting software security, we adopted a two-fold approach. Initially, we performed an empirical study to analyse the perceptions of those who had explored the use of ChatGPT for security tasks and shared their views on Twitter. It was determined that security practitioners view ChatGPT as beneficial for various software security tasks, including vulnerability detection, information retrieval, and penetration testing. Secondly, we designed an experiment aimed at investigating the practicality of this technology when deployed as an oracle in real-world settings. In particular, we focused on vulnerability detection and qualitatively examined ChatGPT outputs for given prompts within this prominent software security task. Based on our analysis, responses from ChatGPT in this task are largely filled with generic security information and may not be appropriate for industry use. To prevent data leakage, we performed this analysis on a vulnerability dataset compiled after the OpenAI data cut-off date from real-world projects covering 40 distinct vulnerability types and 12 programming languages. We assert that the findings from this study would contribute to future research aimed at developing and evaluating LLMs dedicated to software security., Comment: Accepted for publication at International Conference on Trust, Privacy and Security - 2024

We study random constant-depth quantum circuits in a two-dimensional architecture. While these circuits only produce entanglement between nearby qubits on the lattice, long-range entanglement can be generated by measuring a subset of the qubits of the output state. It is conjectured that this long-range measurement-induced entanglement (MIE) proliferates when the circuit depth is at least a constant critical value. For circuits composed of Haar-random two-qubit gates, it is also believed that this coincides with a quantum advantage phase transition in the classical hardness of sampling from the output distribution. Here we provide evidence for a quantum advantage phase transition in the setting of random Clifford circuits. Our work extends the scope of recent separations between the computational power of constant-depth quantum and classical circuits, demonstrating that this kind of advantage is present in canonical random circuit sampling tasks. In particular, we show that in any architecture of random shallow Clifford circuits, the presence of long-range MIE gives rise to an unconditional quantum advantage. In contrast, any depth-d 2D quantum circuit that satisfies a short-range MIE property can be classically simulated efficiently and with depth O(d). Finally, we introduce a two-dimensional, depth-2, "coarse-grained" circuit architecture, composed of random Clifford gates acting on O(log n) qubits, for which we prove the existence of long-range MIE and establish an unconditional quantum advantage., Comment: 25 pages, 5 figures

Printed electronics offer ultra-low manufacturing costs and the potential for on-demand fabrication of flexible hardware. However, significant intrinsic constraints stemming from their large feature sizes and low integration density pose design challenges that hinder their practicality. In this work, we conduct a holistic exploration of printed neural network accelerators, starting from the analog-to-digital interface - a major area and power sink for sensor processing applications - and extending to networks of ternary neurons and their implementation. We propose bespoke ternary neural networks using approximate popcount and popcount-compare units, developed through a multi-phase evolutionary optimization approach and interfaced with sensors via customizable analog-to-binary converters. Our evaluation results show that the presented designs outperform the state of the art, achieving at least 6x improvement in area and 19x in power. To our knowledge, they represent the first open-source digital printed neural network classifiers capable of operating with existing printed energy harvesters., Comment: Accepted for publication at ICCAD '24

This research combines Knowledge Distillation (KD) and Mixture of Experts (MoE) to develop modular, efficient multilingual language models. Key objectives include evaluating adaptive versus fixed alpha methods in KD and comparing modular MoE architectures for handling multi-domain inputs and preventing catastrophic forgetting. KD compresses large language models (LLMs) into smaller, efficient models, while MoE enhances modularity with specialized tasks. Experiments showed similar performance for both KD methods, with marginal improvements from adaptive alpha. A combined loss approach provided more stable learning. The router, trained to classify input sequences into English, French, German, or Python, achieved 99.95% precision, recall, and F1 score, with Logistic Regression being the most effective classifier. Evaluations of modular MoE architectures revealed that Pre-trained Language Experts (PLE) and Joint Expert Embedding Training (JEET) performed similarly, while the MoE with Common Expert (MoE-CE) setup showed slightly lower performance. Including a common expert in MoE-CE improved its performance. Studies on catastrophic forgetting indicated that sequential training led to significant forgetting, while single-session training with balanced batches and the MoE approach mitigated this issue. The MoE architecture preserved knowledge across multiple languages effectively. The research contributes open-sourced resources including the dataset (https://zenodo.org/doi/10.5281/zenodo.12677631), a balanced dataset creation tool (https://github.com/padas-lab-de/multi-language-dataset-creator), and the research codebase (https://github.com/ModMaamari/mixture-modular-experts)., Comment: Preprint

This letter introduces a framework for the automatic generation of hardware cores for Artificial Neural Network (ANN)-based chaotic oscillators. The framework trains the model to approximate a chaotic system, then performs design space exploration yielding potential hardware architectures for its implementation. The framework then generates the corresponding synthesizable High-Level Synthesis code and a validation testbench from a selected solution. The hardware design primarily targets FPGAs. The proposed framework offers a rapid hardware design process of candidate architectures superior to manually designed works in terms of hardware cost and throughput. The source code is available on GitHub., Comment: 5 Pages

Given a smooth positive function $K$ on the standard sphere $(\mathbb{S}^n,g_0)$, we use Morse theoretical methods and counting index formulae to prove that, under generic conditions on the function $K$, there are arbitrarily many metrics $g$ conformally equivalent to $g_0$ and whose scalar curvature is given by the function $K$ provided that the function is sufficiently close to the scalar curvature of $g_0$. Our approach leverages a comprehensive characterization of blowing-up solutions of a subcritical approximation, along with various Morse relations involving their indices. Notably, this multiplicity result is achieved without relying on any symmetry or periodicity assumptions about the function $K$.

This study introduces novel superior scoring rules called Penalized Brier Score (PBS) and Penalized Logarithmic Loss (PLL) to improve model evaluation for probabilistic classification. Traditional scoring rules like Brier Score and Logarithmic Loss sometimes assign better scores to misclassifications in comparison with correct classifications. This discrepancy from the actual preference for rewarding correct classifications can lead to suboptimal model selection. By integrating penalties for misclassifications, PBS and PLL modify traditional proper scoring rules to consistently assign better scores to correct predictions. Formal proofs demonstrate that PBS and PLL satisfy strictly proper scoring rule properties while also preferentially rewarding accurate classifications. Experiments showcase the benefits of using PBS and PLL for model selection, model checkpointing, and early stopping. PBS exhibits a higher negative correlation with the F1 score compared to the Brier Score during training. Thus, PBS more effectively identifies optimal checkpoints and early stopping points, leading to improved F1 scores. Comparative analysis verifies models selected by PBS and PLL achieve superior F1 scores. Therefore, PBS and PLL address the gap between uncertainty quantification and accuracy maximization by encapsulating both proper scoring principles and explicit preference for true classifications. The proposed metrics can enhance model evaluation and selection for reliable probabilistic classification., Comment: 21 Pages, 3 Figures, 3 Tables

Quadratic Programs (QPs) have become a mature technology for the control of robots of all kinds, including humanoid robots. One aspect has been largely overlooked, however, which is the accuracy with which these QPs should be solved. QP solvers aim at providing solutions accurate up to floating point precision ($\approx10^{-8}$). Considering physical quantities expressed in SI or similar units (meters, radians, etc.), such precision seems completely unrelated to both task requirements and hardware capacity. Typically, humanoid robots never achieve, nor are capable of achieving sub-millimeter precision in manipulation tasks. With this observation in mind, our objectives in this paper are two-fold: first examine how the QP solution accuracy impacts the resulting robot motion accuracy, then evaluate how a reduced solution accuracy requirement can be leveraged to reduce the corresponding computational effort. Experiments with a dynamic simulation of RHPS-1 humanoid robot indicate that computational effort can be divided by more than 27 while maintaining the desired motion accuracy., Comment: Submitted to 2024 IEEE-RAS International Conference on Humanoid Robots (Humanoids)

This study investigates the impact of Hofstede's cultural dimensions on abnormal core earnings management in multiple national cultural contexts. We employ an Ordinary Least Squares (OLS) regression model with abnormal core earnings as the dependent variable. The independent variables analyzed include Hofstede's dimensions: Power Distance Index (PDI), Individualism (IDV), Masculinity (MAS), and Uncertainty Avoidance Index (UAI). Our findings reveal that individualism is positively associated with abnormal core earnings, suggesting that cultures characterized by high individualism may encourage practices that inflate earnings due to the prominence of personal achievement and rewards. In contrast, masculinity negatively correlates with abnormal core earnings, indicating that the risk-taking attributes associated with masculine cultures may deter earnings management. Interestingly, uncertainty avoidance is positively linked to abnormal core earnings, supporting the notion that managers tend to engage more in earnings management to minimize fluctuations in financial reports in cultures with high uncertainty avoidance. The relationship between power distance and abnormal core earnings is found to be non-significant, indicating no substantial effect in this context. These findings contribute to the literature on cultural influences in financial reporting, providing valuable insights for policymakers and multinational firms concerning the cultural contexts within which financial decisions and reporting occur., Comment: 12 pages, 1 figure, "Published with European Journal of Business and Management"

Many studies have examined the connection between the intention to start a business and environmental values. However, there still needs to be more knowledge in the extant literature about how climate change campaigns influence sustainable entrepreneurial intention. This study uses the Theory of Planned Behavior (TPB) to develop a theoretical framework to explain how climate change campaigns affect the intention to start a sustainable business. This interdisciplinary conceptual research model bridges the gap between climate awareness, sustainable values, and entrepreneurial intentions, offering a robust framework for understanding and fostering sustainable entrepreneurial behaviors. Our study lays the groundwork for future empirical studies and real-world interventions to advance sustainability through entrepreneurship., Comment: 10 pages, 1 figure, Published with International Research Journal of Economics and Management Studies (IRJEMS)

In the rapidly evolving field of sports analytics, the automation of targeted video processing is a pivotal advancement. We propose PlayerTV, an innovative framework which harnesses state-of-the-art AI technologies for automatic player tracking and identification in soccer videos. By integrating object detection and tracking, Optical Character Recognition (OCR), and color analysis, PlayerTV facilitates the generation of player-specific highlight clips from extensive game footage, significantly reducing the manual labor traditionally associated with such tasks. Preliminary results from the evaluation of our core pipeline, tested on a dataset from the Norwegian Eliteserien league, indicate that PlayerTV can accurately and efficiently identify teams and players, and our interactive Graphical User Interface (GUI) serves as a user-friendly application wrapping this functionality for streamlined use.

Accurate assessment of myocardial tissue stiffness is pivotal for the diagnosis and prognosis of heart diseases. Left ventricular diastolic stiffness ($\beta$) obtained from the end-diastolic pressure-volume relationship (EDPVR) has conventionally been utilized as a representative metric of myocardial stiffness. The EDPVR can be employed to estimate the intrinsic stiffness of myocardial tissues through image-based in-silico inverse optimization. However, whether $\beta$, as an organ-level metric, accurately represents the tissue-level myocardial tissue stiffness in healthy and diseased myocardium remains elusive. We developed a modeling-based approach utilizing a two-parameter material model for the myocardium (denoted by $a_f$ and $b_f$) in image-based in-silico biventricular heart models to generate EDPVRs for different material parameters. Our results indicated a variable relationship between $\beta$ and the material parameters depending on the range of the parameters. Interestingly, $\beta$ showed a very low sensitivity to $a_f$, once averaged across several LV geometries, and even a negative correlation with $a_f$ for small values of $a_f$. These findings call for a critical assessment of the reliability and confoundedness of EDPVR-derived metrics to represent tissue-level myocardial stiffness. Our results also underscore the necessity to explore image-based in-silico frameworks, promising to provide a high-fidelity and potentially non-invasive assessment of myocardial stiffness., Comment: 4 pages, 5 figures, accepted in the IEEE EMBC 2024 conference

Semantic communication (SemCom) leveraging advanced deep learning (DL) technologies enhances the efficiency and reliability of information transmission. Emerging stacked intelligent metasurface (SIM) with an electromagnetic neural network (EMNN) architecture enables complex computations at the speed of light. In this letter, we introduce an innovative SIM-aided SemCom system for image recognition tasks, where a SIM is positioned in front of the transmitting antenna. In contrast to conventional communication systems that transmit modulated signals carrying the image information or compressed semantic information, the carrier EM wave is directly transmitted from the source. The input layer of the SIM performs source encoding, while the remaining multi-layer architecture constitutes an EMNN for semantic encoding, transforming signals into a unique beam towards a receiving antenna corresponding to the image class. Remarkably, both the source and semantic encoding occur naturally as the EM waves propagate through the SIM. At the receiver, the image is recognized by probing the received signal magnitude across the receiving array. To this end, we utilize an efficient mini-batch gradient descent algorithm to train the transmission coefficients of SIM's meta-atoms to learn the semantic representation of the image. Extensive numerical results verify the effectiveness of utilizing the SIM-based EMNN for image recognition task-oriented SemComs, achieving more than 90\% recognition accuracy., Comment: 5 pages, 4 figures, accepted by IEEE Wireless Communications Letters

Lithium-ion batteries are widely used in electric vehicles and grid energy storage systems. Compared to cylindrical batteries, prismatic cells are the primary choice because of their advantage for dense packing. However, thermal runaway and temperature inhomogeneities are the main thermal regulation problems that affect their reliability, safety, and useful life. Here, we propose and assess a multifaceted cooling system composed of water channels (active cooling) and metallic foam embedded with two types of phase-change materials or PCMs (passive cooling) with different melting points. We show that a multifaceted thermal regulation strategy can improve both cooling effectiveness and temperature homogeneity through a representative prismatic battery module. Our numerical results indicate that for a battery pack cooled with a water channel (3C discharge rate), a dual-PCM arrangement can reduce the maximum temperature by 1.3 $^\circ$C and 2.7 $^\circ$C compared to a mono-PCM arrangement and a battery pack without PCM. The maximum temperature difference within the cell is also 1.2 $^\circ$C. Therefore, multi-PCM thermal management systems show better performance than their mono-PCM predecessors in terms of lowering the maximum battery temperature and improving thermal homogeneity. This work motivates the development of multifaceted thermal management systems with active and passive cooling to improve the long-term performance of electrochemical battery cells.