Search

The cellular wireless networks are evolving towards acquiring newer capabilities, such as sensing, which will support novel use cases and applications. Many of these require indoor sensing capabilities, which can be realized by exploiting the perturbation in the indoor channel. In this work, we conduct an indoor channel measurement campaign to study these perturbations and develop AI-based algorithms for estimating sensing parameters. We develop several AI methods based on CNN and tree-based ensemble architectures for sensing. We show that the presence of a passive target like a person can be detected from the channel perturbation of a single link with more than 90 % accuracy with a simple CNN based AI algorithm. However, sensing the position of a passive target is far more challenging requiring more complex AI algorithms and deployments. We show that the position of the human in the indoor room can be estimated within the average position error of 0.7 m with a deployment having three links and employing complex AI architecture for position estimation. We also compare the results with the baseline algorithm to demonstrate the utility of the proposed method., Comment: 4 Pages of technical content and 5th page with reference. Submitted to ICASSP 2025

Automatic Speech Recognition (ASR) systems are pivotal in transcribing speech into text, yet the errors they introduce can significantly degrade the performance of downstream tasks like summarization. This issue is particularly pronounced in clinical dialogue summarization, a low-resource domain where supervised data for fine-tuning is scarce, necessitating the use of ASR models as black-box solutions. Employing conventional data augmentation for enhancing the noise robustness of summarization models is not feasible either due to the unavailability of sufficient medical dialogue audio recordings and corresponding ASR transcripts. To address this challenge, we propose MEDSAGE, an approach for generating synthetic samples for data augmentation using Large Language Models (LLMs). Specifically, we leverage the in-context learning capabilities of LLMs and instruct them to generate ASR-like errors based on a few available medical dialogue examples with audio recordings. Experimental results show that LLMs can effectively model ASR noise, and incorporating this noisy data into the training process significantly improves the robustness and accuracy of medical dialogue summarization systems. This approach addresses the challenges of noisy ASR outputs in critical applications, offering a robust solution to enhance the reliability of clinical dialogue summarization.

Medical abstractive summarization faces the challenge of balancing faithfulness and informativeness. Current methods often sacrifice key information for faithfulness or introduce confabulations when prioritizing informativeness. While recent advancements in techniques like in-context learning (ICL) and fine-tuning have improved medical summarization, they often overlook crucial aspects such as faithfulness and informativeness without considering advanced methods like model reasoning and self-improvement. Moreover, the field lacks a unified benchmark, hindering systematic evaluation due to varied metrics and datasets. This paper addresses these gaps by presenting a comprehensive benchmark of six advanced abstractive summarization methods across three diverse datasets using five standardized metrics. Building on these findings, we propose uMedSum, a modular hybrid summarization framework that introduces novel approaches for sequential confabulation removal followed by key missing information addition, ensuring both faithfulness and informativeness. Our work improves upon previous GPT-4-based state-of-the-art (SOTA) medical summarization methods, significantly outperforming them in both quantitative metrics and qualitative domain expert evaluations. Notably, we achieve an average relative performance improvement of 11.8% in reference-free metrics over the previous SOTA. Doctors prefer uMedSum's summaries 6 times more than previous SOTA in difficult cases where there are chances of confabulations or missing information. These results highlight uMedSum's effectiveness and generalizability across various datasets and metrics, marking a significant advancement in medical summarization., Comment: 12 pages

A hub-based colony consists of multiple agents who share a common nest site called the hub. Agents perform tasks away from the hub like foraging for food or gathering information about future nest sites. Modeling hub-based colonies is challenging because the size of the collective state space grows rapidly as the number of agents grows. This paper presents a graph-based representation of the colony that can be combined with graph-based encoders to create low-dimensional representations of collective state that can scale to many agents for a best-of-N colony problem. We demonstrate how the information in the low-dimensional embedding can be used with two experiments. First, we show how the information in the tensor can be used to cluster collective states by the probability of choosing the best site for a very small problem. Second, we show how structured collective trajectories emerge when a graph encoder is used to learn the low-dimensional embedding, and these trajectories have information that can be used to predict swarm performance.

Model-based evaluation is at the heart of successful model development -- as a reward model for training, and as a replacement for human evaluation. To train such evaluators, the standard approach is to collect a large amount of human preference judgments over model responses, which is costly and the data becomes stale as models improve. In this work, we present an approach that aims to im-prove evaluators without human annotations, using synthetic training data only. Starting from unlabeled instructions, our iterative self-improvement scheme generates contrasting model outputs and trains an LLM-as-a-Judge to produce reasoning traces and final judgments, repeating this training at each new iteration using the improved predictions. Without any labeled preference data, our Self-Taught Evaluator can improve a strong LLM (Llama3-70B-Instruct) from 75.4 to 88.3 (88.7 with majority vote) on RewardBench. This outperforms commonly used LLM judges such as GPT-4 and matches the performance of the top-performing reward models trained with labeled examples.

Generation, statistically steady state, and temporal decay of axially rotating thermal counterflow of superfluid $^4$He (He~II) in a square channel is probed using the second sound attenuation technique, measuring the density of quantized vortex lines. The array of rectilinear quantized vortices created by rotation strongly affects the development of quantum turbulence. At relatively slow angular velocities, the type of instability responsible for the destruction of the laminar counterflow qualitatively changes: the growth of seed vortex loops pinned on the channel wall becomes gradually replaced by the growth due to Donnelly-Glaberson instability, which leads to rapid growth of helical Kelvin waves on vortices parallel with applied counterflow. The initial transient growth of vortex line density that follows the sudden start of the counterflow appears self-similar, linear in dimensionless time, $\Omega t$. We show numerically that Kelvin waves of sufficiently strong amplitude reorient the vortices into more flattened shapes, which grow similarly to a free vortex ring. The observed steady state vortex line density at sufficiently high counterflow velocity and its early temporal decay after the counterflow is switched off is not appreciably affected by rotation. It is striking, however, that although the steady state of rotating counterflow is very different from rotating classical grid-generated turbulence, the late temporal decay of both displays similar features: the decay exponent decreases with the rotation rate $\Omega$ from -3/2 towards approximately -0.7, typical for two-dimensional turbulence, consistent with the transition to bidirectional cascade., Comment: 20 pages, 9 figures, 2 videos

Ordinary 3D Baryon Acoustic Oscillations (BAO) data are model-dependent, requiring the assumption of a cosmological model to calculate comoving distances during data reduction. Throughout the present-day literature, the assumed model is $\Lambda$CDM. However, this assumption can be inadequate when analyzing alternative cosmologies, potentially biasing the Hubble constant ($H_0$) low, thus contributing to the Hubble tension. To address this issue, we use model-independent 2D BAO data and compare the results to those obtained using 3D BAO data. As a concrete example, we investigate the cosmological models provided by bimetric gravity, a modified theory of gravity that transitions from a negative cosmological constant in the early universe to a positive one in the late universe. By combining 2D BAO data with cosmic microwave background and type Ia supernovae data, we find that the inverse distance ladder in this theory yields a Hubble constant of $H_0 = (71.0 \pm 0.9) \, \mathrm{km/s/Mpc}$, consistent with the SH0ES local distance ladder measurement of $H_0 = (73.0 \pm 1.0) \, \mathrm{km/s/Mpc}$. Replacing 2D BAO with 3D BAO results in $H_0 = (68.6 \pm 0.5) \, \mathrm{km/s/Mpc}$ from the inverse distance ladder. We conclude that the choice of BAO data significantly impacts the Hubble tension, with ordinary 3D BAO data exacerbating the tension, while 2D BAO data provides results consistent with the local distance ladder., Comment: v2: extended discussion on relation between bimetric cosmology and other alternative expansion histories

The accurate evaluation of differential treatment in language models to specific groups is critical to ensuring a positive and safe user experience. An ideal evaluation should have the properties of being robust, extendable to new groups or attributes, and being able to capture biases that appear in typical usage (rather than just extreme, rare cases). Relatedly, bias evaluation should surface not only egregious biases but also ones that are subtle and commonplace, such as a likelihood for talking about appearances with regard to women. We present FairPair, an evaluation framework for assessing differential treatment that occurs during ordinary usage. FairPair operates through counterfactual pairs, but crucially, the paired continuations are grounded in the same demographic group, which ensures equivalent comparison. Additionally, unlike prior work, our method factors in the inherent variability that comes from the generation process itself by measuring the sampling variability. We present an evaluation of several commonly used generative models and a qualitative analysis that indicates a preference for discussing family and hobbies with regard to women.

This study analyses the medium access control (MAC) layer aspects of a low-Earth-orbit (LEO) satellite-based Internet of Things (IoT) network. A transmission scheme based on change detection is proposed to accommodate more users within the network and improve energy efficiency. Machine learning (ML) algorithms are also proposed to reduce the payload size by leveraging the correlation among the sensed parameters. Real-world data from an IoT testbed deployed for a smart city application is utilised to analyse the performance regarding collision probability, effective data received and average battery lifetime. The findings reveal that the traffic pattern, post-implementation of the proposed scheme, differs from the commonly assumed Poisson traffic, thus proving the effectiveness of having IoT data from actual deployment. It is demonstrated that the transmission scheme facilitates accommodating more devices while targeting a specific collision probability. Considering the link budget for a direct access NB-IoT scenario, more data is effectively offloaded to the server within the limited visibility of LEO satellites. The average battery lifetimes are also demonstrated to increase by many folds by using the proposed access schemes and ML algorithms.

Quantum nuclear dynamics with wavepacket time-evolution is classically intractable and viewed as a promising avenue for quantum information processing. Here, we use an IonQ 11-qubit trapped-ion quantum computer, Harmony, to study the quantum wavepacket dynamics of a shared-proton within a short-strong hydrogen-bonded system. We also provide the first application of distributed quantum computing for chemical dynamics problems, where the distributed set of quantum processes is constructed using a tensor network formalism. For a range of initial states, we experimentally drive the ion-trap system to emulate the quantum nuclear wavepacket as it evolves along the potential surface generated from electronic structure. Following the experimental creation of the nuclear wavepacket, we extract measurement observables such as its time-dependent spatial projection and its characteristic vibrational frequencies to good agreement with classical results. Vibrational eigenenergies obtained from quantum computational are in agreement with those obtained from classical simulations to within a fraction of a kcal/mol, thus suggesting chemical accuracy. Our approach opens a new paradigm for studying the quantum chemical dynamics and vibrational spectra of molecules and also provides the first demonstration for parallel quantum computation on a distributed set of ion-trap quantum computers.

Information Retrieval (IR) systems are designed to deliver relevant content, but traditional systems may not optimize rankings for fairness, neutrality, or the balance of ideas. Consequently, IR can often introduce indexical biases, or biases in the positional order of documents. Although indexical bias can demonstrably affect people's opinion, voting patterns, and other behaviors, these issues remain understudied as the field lacks reliable metrics and procedures for automatically measuring indexical bias. Towards this end, we introduce the PAIR framework, which supports automatic bias audits for ranked documents or entire IR systems. After introducing DUO, the first general-purpose automatic bias metric, we run an extensive evaluation of 8 IR systems on a new corpus of 32k synthetic and 4.7k natural documents, with 4k queries spanning 1.4k controversial issue topics. A human behavioral study validates our approach, showing that our bias metric can help predict when and how indexical bias will shift a reader's opinion., Comment: ACL 2024

In this article we report the structural stability of Er$_2$Ti$_2$O$_7$ cubic pyrochlore with pressure using x-ray diffraction, Raman spectroscopy, photoluminescence, x-ray absorption and ab-initio calculations. Our studies establish a phase transformation in Er$_2$Ti$_2$O$_7$ from ambient cubic phase to high-pressure orthorhombic (cotunnite) phase, initiated at ~40 GPa. The transformation is sluggish and it does not complete even at the highest measured pressure in our study i.e. ~60.0 GPa. This is further supported by the first principle calculations which reveal that cotunnite phase is energetically more stable than the ambient phase above ~53 GPa. After complete release of pressure, the high-pressure cotunnite phase is retained while the fraction of untransformed pyrochlore phase becomes amorphous. Furthermore, the EXAFS data of the recovered sample at L3 edge of Er3+ ion show an increase in the coordination number of cations from eight at ambient to nine in the high-pressure phase. The mechanism of structural transformation is explained in terms of accumulation of cation antisite defects and subsequent disordering of cations and anions in their respective sublattice. The amorphization of the pyrochlore phase upon release is interpreted as the inability of accommodating the point defects at ambient conditions, which are formed in the pyrochlore lattice under compression.

There is vivid research on adapting Large Language Models (LLMs) to perform a variety of tasks in high-stakes domains such as healthcare. Despite their popularity, there is a lack of understanding of the extent and contributing factors that allow LLMs to recall relevant knowledge and combine it with presented information in the clinical and biomedical domain: a fundamental pre-requisite for success on down-stream tasks. Addressing this gap, we use Multiple Choice and Abstractive Question Answering to conduct a large-scale empirical study on 22 datasets in three generalist and three specialist biomedical sub-domains. Our multifaceted analysis of the performance of 15 LLMs, further broken down by sub-domain, source of knowledge and model architecture, uncovers success factors such as instruction tuning that lead to improved recall and comprehension. We further show that while recently proposed domain-adapted models may lack adequate knowledge, directly fine-tuning on our collected medical knowledge datasets shows encouraging results, even generalising to unseen specialist sub-domains. We complement the quantitative results with a skill-oriented manual error analysis, which reveals a significant gap between the models' capabilities to simply recall necessary knowledge and to integrate it with the presented context. To foster research and collaboration in this field we share M-QALM, our resources, standardised methodology, and evaluation results, with the research community to facilitate further advancements in clinical knowledge representation learning within language models., Comment: Accepted at ACL 2024 (Findings)

The main purpose of this paper is to design a local discontinuous Galerkin (LDG) method for the Benjamin-Ono equation. We analyze the stability and error estimates for the semi-discrete LDG scheme. We prove that the scheme is $L^2$-stable and it converges at a rate $\mathcal{O}(h^{k+1/2})$ for general nonlinear flux. Furthermore, we develop a fully discrete LDG scheme using the four-stage fourth order Runge-Kutta method and ensure the devised scheme is strongly stable in case of linear flux using two-step and three-step stability approach under an appropriate time step constraint. Numerical examples are provided to validate the efficiency and accuracy of the method., Comment: arXiv admin note: text overlap with arXiv:2404.18069

In this paper, the authors propose the utilization of Fibonacci Neural Networks (FNN) for solving arbitrary order differential equations. The FNN architecture comprises input, middle, and output layers, with various degrees of Fibonacci polynomials serving as activation functions in the middle layer. The trial solution of the differential equation is treated as the output of the FNN, which involves adjustable parameters (weights). These weights are iteratively updated during the training of the Fibonacci neural network using backpropagation. The efficacy of the proposed method is evaluated by solving five differential problems with known exact solutions, allowing for an assessment of its accuracy. Comparative analyses are conducted against previously established techniques, demonstrating superior accuracy and efficacy in solving the addressed problems.

The widespread adoption of blockchain technology has amplified the spectrum of potential threats to its integrity and security. The ongoing quest to exploit vulnerabilities emphasizes how critical it is to expand on current research initiatives. Thus, using a methodology based on discrete blockchain layers, our survey study aims to broaden the existing body of knowledge by thoroughly discussing both new and known attack vectors inside the blockchain ecosystem. This survey proposes a novel classification of blockchain attacks and an in-depth investigation of blockchain data security. In particular, the paper provides a thorough discussion of the attack techniques and vulnerabilities that are specific to each tier, along with a detailed look at mitigating techniques. We reveal the deep dynamics of these security concerns by closely investigating the fundamental causes of attacks at various blockchain tiers. We clarify mitigation methods for known vulnerabilities and offer new information on recently developed attack vectors. We also discuss the implications of quantum computing in blockchain and the weaknesses in the current technology that can be exploited in the future. Our study advances the field of blockchain security and privacy research while also contributing to our understanding of blockchain vulnerabilities and attacks. This survey paper is a useful tool for readers who want to learn more about the intricacies of blockchain security. It also invites researchers to help strengthen blockchain privacy and security, paving the way for further developments in this dynamic and ever-evolving field.

We propose a local discontinuous Galerkin (LDG) method for fractional Korteweg-de Vries equation involving the fractional Laplacian with exponent $\alpha\in (1,2)$ in one and two space dimensions. By decomposing the fractional Laplacian into a first order derivative and a fractional integral, we prove $L^2$-stability of the semi-discrete LDG scheme incorporating suitable interface and boundary fluxes. We analyze the error estimate by considering linear convection term and utilizing the estimate, we derive the error estimate for general nonlinear flux and demonstrate an order of convergence $\mathcal{O}(h^{k+1/2})$. Moreover, the stability and error analysis have been extended to multiple space dimensional case. Numerical illustrations are shown to demonstrate the efficiency of the scheme by obtaining an optimal order of convergence.

We address the problem of regressing 3D human pose and shape from a single image, with a focus on 3D accuracy. The current best methods leverage large datasets of 3D pseudo-ground-truth (p-GT) and 2D keypoints, leading to robust performance. With such methods, we observe a paradoxical decline in 3D pose accuracy with increasing 2D accuracy. This is caused by biases in the p-GT and the use of an approximate camera projection model. We quantify the error induced by current camera models and show that fitting 2D keypoints and p-GT accurately causes incorrect 3D poses. Our analysis defines the invalid distances within which minimizing 2D and p-GT losses is detrimental. We use this to formulate a new loss Threshold-Adaptive Loss Scaling (TALS) that penalizes gross 2D and p-GT losses but not smaller ones. With such a loss, there are many 3D poses that could equally explain the 2D evidence. To reduce this ambiguity we need a prior over valid human poses but such priors can introduce unwanted bias. To address this, we exploit a tokenized representation of human pose and reformulate the problem as token prediction. This restricts the estimated poses to the space of valid poses, effectively providing a uniform prior. Extensive experiments on the EMDB and 3DPW datasets show that our reformulated keypoint loss and tokenization allows us to train on in-the-wild data while improving 3D accuracy over the state-of-the-art. Our models and code are available for research at https://tokenhmr.is.tue.mpg.de., Comment: CVPR 2024

Interactions among biologically active agents is facilitated by their self-generated chemical and hydrodynamic fields. In order to elucidate the pair-wise interactions between such micro-organisms, we employ active droplets as a model system, capable of self-generating chemical and hydrodynamic fields. We demonstrate that the solute P\'eclet number ($Pe$), characterizing the relative strength of its convective to diffusive transport, plays a crucial role in determining how the chemical and hydrodynamic fields impact their interactions. Our findings reveal that at low $Pe$, the interaction is predominantly governed by chemo-repulsive effects, leading to droplets avoiding physical contact. Conversely, at elevated $Pe$, hydrodynamic interactions become more influential, leading to physical engagement. However, irrespective of $Pe$, the interaction of a droplet with the chemical trail of another droplet is always governed by chemo-repulsive effects. Furthermore, our results establish that the chemo-repulsive deflection/rebounding of droplets is influenced by the droplets' inherent chemical polarity, as determined by its $Pe$, independent of their approach orientation. Our findings offer a methodology for tuning the outcomes of binary interactions among chemically active droplets, laying the groundwork for potential studies on their collective dynamics., Comment: 11 pages, 6 figures

Modern compression methods can summarize a target distribution $\mathbb{P}$ more succinctly than i.i.d. sampling but require access to a low-bias input sequence like a Markov chain converging quickly to $\mathbb{P}$. We introduce a new suite of compression methods suitable for compression with biased input sequences. Given $n$ points targeting the wrong distribution and quadratic time, Stein kernel thinning (SKT) returns $\sqrt{n}$ equal-weighted points with $\widetilde{O}(n^{-1/2})$ maximum mean discrepancy (MMD) to $\mathbb{P}$. For larger-scale compression tasks, low-rank SKT achieves the same feat in sub-quadratic time using an adaptive low-rank debiasing procedure that may be of independent interest. For downstream tasks that support simplex or constant-preserving weights, Stein recombination and Stein Cholesky achieve even greater parsimony, matching the guarantees of SKT with as few as $\text{poly-log}(n)$ weighted points. Underlying these advances are new guarantees for the quality of simplex-weighted coresets, the spectral decay of kernel matrices, and the covering numbers of Stein kernel Hilbert spaces. In our experiments, our techniques provide succinct and accurate posterior summaries while overcoming biases due to burn-in, approximate Markov chain Monte Carlo, and tempering., Comment: Published at ICML 2024

Electric vehicles (EVs) are increasingly becoming popular as a viable means of transportation for the future. The use of EVs may help to provide better climatic conditions in urban areas with a pocket friendly cost for transportation to the consumers throughout its life. EVs enact as a boon to the society by providing zero tailpipe emissions, better comfort, low lifecycle cost and higher connectivity. The article aims to provide scientific information through the literature across various aspects of EVs in their lifetime and thus, assist the scholarly community and various organisations to understand the impact of EVs. In this study we have gathered information from the articles published in SCOPUS database and through grey literature with the focus of information post 2009. After identification of various factors while purchasing the vehicle, the total cost of ownership (TCO) for vehicles is calculated and their average TCO for each segment is considered for the study. Following that, we investigate the ranking of EVs, vehicles powered by internal combustion engines (ICEVs), and hybrid electric vehicles (HEVs) in a variety of price segments by employing a combination of two different multi-criteria decision making (MCDM) techniques. Initially, best-worst method (BWM) is used to determine the weights of each of the identified criterion, which is thereafter, used in conjunction with the technique for order preference by similarity to ideal solution (TOPSIS) to compute the rank of the available alternatives using BWM. The ranking obtained clearly indicates that EVs should be first purchase choice of the consumers, followed by HEVs and ICEVs respectively. Thus, the results help us conclude that EVs enact as a sustainable means of transport for the future., Comment: 25 pages, 3 figures, 3 tables

With the advent of social media, fun selfie filters have come into tremendous mainstream use affecting the functioning of facial biometric systems as well as image recognition systems. These filters vary from beautification filters and Augmented Reality (AR)-based filters to filters that modify facial landmarks. Hence, there is a need to assess the impact of such filters on the performance of existing face recognition systems. The limitation associated with existing solutions is that these solutions focus more on the beautification filters. However, the current AR-based filters and filters which distort facial key points are in vogue recently and make the faces highly unrecognizable even to the naked eye. Also, the filters considered are mostly obsolete with limited variations. To mitigate these limitations, we aim to perform a holistic impact analysis of the latest filters and propose an user recognition model with the filtered images. We have utilized a benchmark dataset for baseline images, and applied the latest filters over them to generate a beautified/filtered dataset. Next, we have introduced a model FaceFilterNet for beautified user recognition. In this framework, we also utilize our model to comment on various attributes of the person including age, gender, and ethnicity. In addition, we have also presented a filter-wise impact analysis on face recognition, age estimation, gender, and ethnicity prediction. The proposed method affirms the efficacy of our dataset with an accuracy of 87.25% and an optimal accuracy for facial attribute analysis.

Given the destructive impacts of tropical cyclones, it is critical to have a reliable system for cyclone intensity detection. Various techniques are available for this purpose, each with differing levels of accuracy. In this paper, we introduce two ensemble-based models based on AlexNet architecture to estimate tropical cyclone intensity using visible satellite images. The first model, trained on the entire dataset, is called the global AlexNet model. The second model is a distributed version of AlexNet in which multiple AlexNets are trained separately on subsets of the training data categorized according to the Saffir-Simpson wind speed scale prescribed by the meterologists. We evaluated the performance of both models against a deep learning benchmark model called \textit{Deepti} using a publicly available cyclone image dataset. Results indicate that both the global model (with a root mean square error (RMSE) of 9.03 knots) and the distributed model (with a RMSE of 9.3 knots) outperform the benchmark model (with a RMSE of 13.62 knots). We provide a thorough discussion of our solution approach, including an explanantion of the AlexNet's performance using gradient class activation maps (grad-CAM). Our proposed solution strategy allows future experimentation with various deep learning models in both single and multi-channel settings.

We compute the $\text{PSL}(2,\mathbb{C})$ Chern-Simons partition function of a closed 3-manifold obtained from Dehn fillings of the link complement $\mathbf S^3\backslash {\mathcal{L}}$, where $\mathcal{L}=\mathcal{K}# H$ is the connected sum of the knot $\mathcal {K}$ with the Hopf link $H$. Motivated by our earlier work on topological entanglement and the reduced density matrix $\sigma$ for such link complements, we wanted to determine a choice of Dehn filling so that the trace of the matrix $\sigma$ becomes equal to the $\text{PSL}(2,\mathbb{C})$ partition function of the closed 3-manifold. We use the SnapPy program and numerical techniques to show this equivalence up to the leading order. We have given explicit results for all hyperbolic knots $\mathcal{K}$ up to six crossings., Comment: 25 pages, 5 figures, 5 tables

We formulate and study the negative gradient flow of an energy functional of Spin(7)-structures on compact 8-manifolds. The energy functional is the $L^2$-norm of the torsion of the Spin(7)-structure. Our main result is the short-time existence and uniqueness of solutions to the flow. We also explain how this negative gradient flow is the most general flow of Spin(7)-structures. We also study solitons of the flow and prove a non-existence result for compact expanding solitons., Comment: 23 pages. Comments welcome

We present the formation of quasi-periodic cool spicule-like jets in the solar atmosphere using 2.5-D numerical simulation in two-fluid regime (ions+neutrals) under the presence of thermal conduction and ion-neutral collision. The non-linear, impulsive Alfv\'enic perturbations at the top of the photosphere trigger field aligned magnetoacoustic perturbations due to ponderomotive force. The transport of energy from Alfv\'en pulse to such vertical velocity perturbations due to ponderomotive force is considered as an initial trigger mechanism. Thereafter, these velocity perturbations steepen into the shocks followed by quasi-periodic rise and fall of the cool jets transporting mass in the overlying corona., Comment: In Press; Philosophical Transactions of the Royal Society A; 21 Pages; 06 Figures. Philosophical Transactions is the oldest English science journal in the world, which has been published continuously since March 1665 as launched by Henry Oldenburg

In this paper, we present and analyze fully discrete finite difference schemes designed for solving the initial value problem associated with the fractional Korteweg-de Vries (KdV) equation involving the fractional Laplacian. We design the scheme by introducing the discrete fractional Laplacian operator which is consistent with the continuous operator, and posses certain properties which are instrumental for the convergence analysis. Assuming the initial data (u_0 \in H^{1+\alpha}(\mathbb{R})), where (\alpha \in [1,2)), our study establishes the convergence of the approximate solutions obtained by the fully discrete finite difference schemes to a classical solution of the fractional KdV equation. Theoretical results are validated through several numerical illustrations for various values of fractional exponent $\alpha$. Furthermore, we demonstrate that the Crank-Nicolson finite difference scheme preserves the inherent conserved quantities along with the improved convergence rates.

Faults on electrical power lines could severely compromise both the reliability and safety of power systems, leading to unstable power delivery and increased outage risks. They pose significant safety hazards, necessitating swift detection and mitigation to maintain electrical infrastructure integrity and ensure continuous power supply. Hence, accurate detection and categorization of electrical faults are pivotal for optimized power system maintenance and operation. In this work, we propose a novel approach for detecting and categorizing electrical faults using the Hessian locally linear embedding (HLLE) technique and subsequent clustering with t-SNE (t-distributed stochastic neighbor embedding) and Gaussian mixture model (GMM). First, we employ HLLE to transform high-dimensional (HD) electrical data into low-dimensional (LD) embedding coordinates. This technique effectively captures the inherent variations and patterns in the data, enabling robust feature extraction. Next, we perform the Mann-Whitney U test based on the feature space of the embedding coordinates for fault detection. This statistical approach allows us to detect electrical faults providing an efficient means of system monitoring and control. Furthermore, to enhance fault categorization, we employ t-SNE with GMM to cluster the detected faults into various categories. To evaluate the performance of the proposed method, we conduct extensive simulations on an electrical system integrated with solar farm. Our results demonstrate that the proposed approach exhibits effective fault detection and clustering across a range of fault types with different variations of the same fault. Overall, this research presents an effective methodology for robust fault detection and categorization in electrical systems, contributing to the advancement of fault management practices and the prevention of system failures.

Electrical fault classification is vital for ensuring the reliability and safety of power systems. Accurate and efficient fault classification methods are essential for timely and effective maintenance. In this paper, we propose a novel approach for effective fault classification through Grassmann manifolds, which is a non-Euclidean space that captures the intrinsic structure of high-dimensional data and offers a robust framework for feature extraction. We use simulated data for electrical distribution systems with various types of electrical faults. The proposed method involves transforming the measurement fault data into Grassmann manifold space using techniques from differential geometry. This transformation aids in uncovering the underlying fault patterns and reducing the computational complexity of subsequent classification steps. To achieve fault classification, we employ a machine learning technique optimized for the Grassmann manifold. The support vector machine classifier is adapted to operate within the Grassmann manifold space, enabling effective discrimination between different fault classes. The results illustrate the efficacy of the proposed Grassmann manifold-based approach for electrical fault classification which showcases its ability to accurately differentiate between various fault types.

Reinforcement Learning from Human Feedback (\textbf{RLHF}) has emerged as a dominant approach for aligning LLM outputs with human preferences. Inspired by the success of RLHF, we study the performance of multiple algorithms that learn from feedback (Expert Iteration, Proximal Policy Optimization (\textbf{PPO}), Return-Conditioned RL) on improving LLM reasoning capabilities. We investigate both sparse and dense rewards provided to the LLM both heuristically and via a learned reward model. We additionally start from multiple model sizes and initializations both with and without supervised fine-tuning (\textbf{SFT}) data. Overall, we find all algorithms perform comparably, with Expert Iteration performing best in most cases. Surprisingly, we find the sample complexity of Expert Iteration is similar to that of PPO, requiring at most on the order of $10^6$ samples to converge from a pretrained checkpoint. We investigate why this is the case, concluding that during RL training models fail to explore significantly beyond solutions already produced by SFT models. Additionally, we discuss a trade off between maj@1 and pass@96 metric performance during SFT training and how conversely RL training improves both simultaneously. We then conclude by discussing the implications of our findings for RLHF and the future role of RL in LLM fine-tuning.

In this paper, we study the Appell function $F_4$ from discrete point of view. In particular, we obtain regions of convergence, difference-differential equations, finite and infinite summation formulas and a list of recursion relations satisfied by the discrete analogues of Appell function $F_4$.

Emphasizing problem formulation in AI literacy activities with children is vital, yet we lack empirical studies on their structure and affordances. We propose that participatory design involving teachable machines facilitates problem formulation activities. To test this, we integrated problem reduction heuristics into storyboarding and invited a university-based intergenerational design team of 10 children (ages 8-13) and 9 adults to co-design a teachable machine. We find that children draw from personal experiences when formulating AI problems; they assume voice and video capabilities, explore diverse machine learning approaches, and plan for error handling. Their ideas promote human involvement in AI, though some are drawn to more autonomous systems. Their designs prioritize values like capability, logic, helpfulness, responsibility, and obedience, and a preference for a comfortable life, family security, inner harmony, and excitement as end-states. We conclude by discussing how these results can inform the design of future participatory AI activities.

Dense retrievers compress source documents into (possibly lossy) vector representations, yet there is little analysis of what information is lost versus preserved, and how it affects downstream tasks. We conduct the first analysis of the information captured by dense retrievers compared to the language models they are based on (e.g., BERT versus Contriever). We use 25 MultiBert checkpoints as randomized initialisations to train MultiContrievers, a set of 25 contriever models. We test whether specific pieces of information -- such as gender and occupation -- can be extracted from contriever vectors of wikipedia-like documents. We measure this extractability via information theoretic probing. We then examine the relationship of extractability to performance and gender bias, as well as the sensitivity of these results to many random initialisations and data shuffles. We find that (1) contriever models have significantly increased extractability, but extractability usually correlates poorly with benchmark performance 2) gender bias is present, but is not caused by the contriever representations 3) there is high sensitivity to both random initialisation and to data shuffle, suggesting that future retrieval research should test across a wider spread of both.

Teaching language models to use tools is an important milestone towards building general assistants, but remains an open problem. While there has been significant progress on learning to use specific tools via fine-tuning, language models still struggle with learning how to robustly use new tools from only a few demonstrations. In this work we introduce a self-verification method which distinguishes between close candidates by self-asking contrastive questions during (1) tool selection; and (2) parameter generation. We construct synthetic, high-quality, self-generated data for this goal using Llama-2 70B, which we intend to release publicly. Extensive experiments on 4 tasks from the ToolBench benchmark, consisting of 17 unseen tools, demonstrate an average improvement of 22% over few-shot baselines, even in scenarios where the distinctions between candidate tools are finely nuanced.

We construct explicit pseudorandom generators that fool $n$-variate polynomials of degree at most $d$ over a finite field $\mathbb{F}_q$. The seed length of our generators is $O(d \log n + \log q)$, over fields of size exponential in $d$ and characteristic at least $d(d-1)+1$. Previous constructions such as Bogdanov's (STOC 2005) and Derksen and Viola's (FOCS 2022) had either suboptimal seed length or required the field size to depend on $n$. Our approach follows Bogdanov's paradigm while incorporating techniques from Lecerf's factorization algorithm (J. Symb. Comput. 2007) and insights from the construction of Derksen and Viola regarding the role of indecomposability of polynomials.

This article introduces a new estimator of average treatment effects under unobserved confounding in modern data-rich environments featuring large numbers of units and outcomes. The proposed estimator is doubly robust, combining outcome imputation, inverse probability weighting, and a novel cross-fitting procedure for matrix completion. We derive finite-sample and asymptotic guarantees, and show that the error of the new estimator converges to a mean-zero Gaussian distribution at a parametric rate. Simulation results demonstrate the practical relevance of the formal properties of the estimators analyzed in this article.

State-of-the-art language models can exhibit impressive reasoning refinement capabilities on math, science or coding tasks. However, recent work demonstrates that even the best models struggle to identify \textit{when and where to refine} without access to external feedback. Outcome-based Reward Models (\textbf{ORMs}), trained to predict correctness of the final answer indicating when to refine, offer one convenient solution for deciding when to refine. Process Based Reward Models (\textbf{PRMs}), trained to predict correctness of intermediate steps, can then be used to indicate where to refine. But they are expensive to train, requiring extensive human annotations. In this paper, we propose Stepwise ORMs (\textbf{SORMs}) which are trained, only on synthetic data, to approximate the expected future reward of the optimal policy or $V^{\star}$. More specifically, SORMs are trained to predict the correctness of the final answer when sampling the current policy many times (rather than only once as in the case of ORMs). Our experiments show that SORMs can more accurately detect incorrect reasoning steps compared to ORMs, thus improving downstream accuracy when doing refinements. We then train \textit{global} refinement models, which take only the question and a draft solution as input and predict a corrected solution, and \textit{local} refinement models which also take as input a critique indicating the location of the first reasoning error. We generate training data for both models synthetically by reusing data used to train the SORM. We find combining global and local refinements, using the ORM as a reranker, significantly outperforms either one individually, as well as a best of three sample baseline. With this strategy we can improve the accuracy of a LLaMA-2 13B model (already fine-tuned with RL) on GSM8K from 53\% to 65\% when greedily sampled.

The introduction of AI and ML technologies into medical devices has revolutionized healthcare diagnostics and treatments. Medical device manufacturers are keen to maximize the advantages afforded by AI and ML by consolidating multiple applications onto a single platform. However, concurrent execution of several AI applications, each with its own visualization components, leads to unpredictable end-to-end latency, primarily due to GPU resource contentions. To mitigate this, manufacturers typically deploy separate workstations for distinct AI applications, thereby increasing financial, energy, and maintenance costs. This paper addresses these challenges within the context of NVIDIA's Holoscan platform, a real-time AI system for streaming sensor data and images. We propose a system design optimized for heterogeneous GPU workloads, encompassing both compute and graphics tasks. Our design leverages CUDA MPS for spatial partitioning of compute workloads and isolates compute and graphics processing onto separate GPUs. We demonstrate significant performance improvements across various end-to-end latency determinism metrics through empirical evaluation with real-world Holoscan medical device applications. For instance, the proposed design reduces maximum latency by 21-30% and improves latency distribution flatness by 17-25% for up to five concurrent endoscopy tool tracking AI applications, compared to a single-GPU baseline. Against a default multi-GPU setup, our optimizations decrease maximum latency by 35% for up to six concurrent applications by improving GPU utilization by 42%. This paper provides clear design insights for AI applications in the edge-computing domain including medical systems, where performance predictability of concurrent and heterogeneous GPU workloads is a critical requirement.

To achieve faithful reasoning that aligns with human expectations, large language models (LLMs) need to ground their reasoning to real-world knowledge (e.g., web facts, math and physical rules). Tools help LLMs access this external knowledge, but there remains challenges for fine-tuning LLM agents (e.g., Toolformer) to invoke tools in multi-step reasoning problems, where inter-connected tool calls require holistic and efficient tool usage planning. In this work, we propose a new method for LLMs to better leverage tools in multi-step reasoning. Our method, Chain-of-Abstraction (CoA), trains LLMs to first decode reasoning chains with abstract placeholders, and then call domain tools to reify each reasoning chain by filling in specific knowledge. This planning with abstract chains enables LLMs to learn more general reasoning strategies, which are robust to shifts of domain knowledge (e.g., math results) relevant to different reasoning questions. It also allows LLMs to perform decoding and calling of external tools in parallel, which avoids the inference delay caused by waiting for tool responses. In mathematical reasoning and Wiki QA domains, we show that our method consistently outperforms previous chain-of-thought and tool-augmented baselines on both in-distribution and out-of-distribution test sets, with an average ~6% absolute QA accuracy improvement. LLM agents trained with our method also show more efficient tool use, with inference speed being on average ~1.4x faster than baseline tool-augmented LLMs.

Possible reasons for the non-detection of absorption in the metastable HeI(2^3S) line at transit observations of warm Neptune GJ436b, in spite of the well pronounced strong absorption features measured earlier in Ly{\alpha} for this planet, are investigated. We perform numeric simulations of the escaping upper atmosphere of this planet and its HeI(2^3S) triplet absorption with a global 3D multi-fluid self-consistent hydrodynamic model. By fitting the model parameters to the lowest detection level of absorption measurements, we constrain an upper limit the He/H abundance three times smaller than the solar value. We demonstrate that neither the significant changes of the stellar wind related with possible stellar coronal mass ejections (CMEs), or possible variations in the stellar ionization radiation, nor the presence of heavy trace elements have crucial effect on the absorption at the 10830{\AA} line of HeI(2^3S) triplet. The main reason of weak signature is that the region populated by the absorbing metastable helium is rather small (<3R_p), as well as the small size of the planet itself, in comparison to the host star. We show that the radiation pressure force acting on the HeI(2^3S) atoms spreads them along the line of sight and around the planet, thus further reducing peak absorption., Comment: 17 pages, 7 figures

Self-propelled droplets serve as ideal model systems to delve deeper into understanding of the motion of biological micro-swimmers by simulating their motility. Biological microorganisms are renowned for showcasing a diverse array of dynamic swimming behaviors when confronted with physical constraints. This study aims to elucidate the impact of physical constraints on swimming characteristics of biological microorganisms. To achieve this, we present observations on the individual and pair-wise behavior of micellar solubilized self-propelled 4-Cyano-4'-pentyl-biphenyl (5CB) oil droplets in a square capillary channel filled with a surfactant trimethyl ammonium bromide (TTAB) aqueous solution. To explore the effect of the underlying P\'eclet ($Pe$) number of the swimming droplets, the study is also performed in the presence of additives such as high molecular weight polymer Polyethylene oxide (PEO) and molecular solute glycerol. The capillary confinement restricts droplet to predominantly one-dimensional (1D) motion, albeit with noticeable differences in their motion across the three scenarios. Through a characterization of the chemical and hydrodynamic flow fields surrounding the droplets, we illustrate that the modification of the droplets' chemical field due to confinement varies significantly based on the underlying differences in the P\'eclet number ($Pe$) in these cases. This alteration in the chemical field distribution notably affects the individual droplets' motion. Moreover, these distinct chemical field interactions between the droplets also lead to variations in their pair-wise motion, ranging from behaviors like chasing to scattering., Comment: 13 pages, 9 figures