Search

Realistic and interactive surgical simulation has the potential to facilitate crucial applications, such as medical professional training and autonomous surgical agent training. In the natural visual domain, world models have enabled action-controlled data generation, demonstrating the potential to train autonomous agents in interactive simulated environments when large-scale real data acquisition is infeasible. However, such works in the surgical domain have been limited to simplified computer simulations, and lack realism. Furthermore, existing literature in world models has predominantly dealt with action-labeled data, limiting their applicability to real-world surgical data, where obtaining action annotation is prohibitively expensive. Inspired by the recent success of Genie in leveraging unlabeled video game data to infer latent actions and enable action-controlled data generation, we propose the first surgical vision world model. The proposed model can generate action-controllable surgical data and the architecture design is verified with extensive experiments on the unlabeled SurgToolLoc-2022 dataset. Codes and implementation details are available at https://github.com/bhattarailab/Surgical-Vision-World-Model

Optimal allocation of scarce resources is a common problem for decision makers faced with choosing a limited number of locations for intervention. Spatiotemporal prediction models could make such decisions data-driven. A recent performance metric called fraction of best possible reach (BPR) measures the impact of using a model's recommended size K subset of sites compared to the best possible top-K in hindsight. We tackle two open problems related to BPR. First, we explore how to rank all sites numerically given a probabilistic model that predicts event counts jointly across sites. Ranking via the per-site mean is suboptimal for BPR. Instead, we offer a better ranking for BPR backed by decision theory. Second, we explore how to train a probabilistic model's parameters to maximize BPR. Discrete selection of K sites implies all-zero parameter gradients which prevent standard gradient training. We overcome this barrier via advances in perturbed optimizers. We further suggest a training objective that combines likelihood with a decision-aware BPR constraint to deliver high-quality top-K rankings as well as good forecasts for all sites. We demonstrate our approach on two where-to-intervene applications: mitigating opioid-related fatal overdoses for public health and monitoring endangered wildlife., Comment: 9 pages, 3 figures

Recent advances in multimodal Human-Robot Interaction (HRI) datasets emphasize the integration of speech and gestures, allowing robots to absorb explicit knowledge and tacit understanding. However, existing datasets primarily focus on elementary tasks like object pointing and pushing, limiting their applicability to complex domains. They prioritize simpler human command data but place less emphasis on training robots to correctly interpret tasks and respond appropriately. To address these gaps, we present the NatSGLD dataset, which was collected using a Wizard of Oz (WoZ) method, where participants interacted with a robot they believed to be autonomous. NatSGLD records humans' multimodal commands (speech and gestures), each paired with a demonstration trajectory and a Linear Temporal Logic (LTL) formula that provides a ground-truth interpretation of the commanded tasks. This dataset serves as a foundational resource for research at the intersection of HRI and machine learning. By providing multimodal inputs and detailed annotations, NatSGLD enables exploration in areas such as multimodal instruction following, plan recognition, and human-advisable reinforcement learning from demonstrations. We release the dataset and code under the MIT License at https://www.snehesh.com/natsgld/ to support future HRI research., Comment: arXiv admin note: substantial text overlap with arXiv:2403.02274

There has been significant prior work using templates to study bias against demographic attributes in MLMs. However, these have limitations: they overlook random variability of templates and target concepts analyzed, assume equality amongst templates, and overlook bias quantification. Addressing these, we propose a systematic statistical approach to assess bias in MLMs, using mixed models to account for random effects, pseudo-perplexity weights for sentences derived from templates and quantify bias using statistical effect sizes. Replicating prior studies, we match on bias scores in magnitude and direction with small to medium effect sizes. Next, we explore the novel problem of gender bias in the context of $\textit{personality}$ and $\textit{character}$ traits, across seven MLMs (base and large). We find that MLMs vary; ALBERT is unbiased for binary gender but the most biased for non-binary $\textit{neo}$, while RoBERTa-large is the most biased for binary gender but shows small to no bias for $\textit{neo}$. There is some alignment of MLM bias and findings in psychology (human perspective) - in $\textit{agreeableness}$ with RoBERTa-large and $\textit{emotional stability}$ with BERT-large. There is general agreement for the remaining 3 personality dimensions: both sides observe at most small differences across gender. For character traits, human studies on gender bias are limited thus comparisons are not feasible., Comment: To appear at Findings of NAACL 2025

Optical crosstalk from sub-milliwatt classical-channel power into quantum channels presents a significant challenge in quantum network development, introducing substantial noise that limits the network's performance, scalability, and fidelity. Here we report a demonstration using photon counting optical time-domain reflectometry ({\nu}-OTDR) to precisely identify and localize crosstalk between separate channels within the same fiber and between separate fibers. The coexistence of classical and quantum signals in the same network necessitates the use of optical switches for efficient routing and control. Crosstalk characterization of an optical switch reveals that crosstalk depends strongly on cross connect configuration, with higher levels observed when connections are presumed to be physically closer and lower levels when further apart. Additionally, we found that crosstalk exhibits a pronounced wavelength dependence, increasing over tenfold at longer wavelengths. These findings demonstrate the value of {\nu}-OTDR in diagnosing and mitigating crosstalk in quantum networks. They highlight the importance of optimizing optical switch configurations and wavelength management to minimize noise, ultimately enhancing the scalability, fidelity, and overall performance of quantum networks. This work establishes a foundational approach to addressing crosstalk, paving the way for more robust and efficient quantum network designs.

Embeddings serve as condensed vector representations for real-world entities, finding applications in Natural Language Processing (NLP), Computer Vision, and Data Management across diverse downstream tasks. Here, we introduce novel specialized embeddings optimized, and explicitly tailored to encode the intricacies of complex 2-D context in tables, featuring horizontal, vertical hierarchical metadata, and nesting. To accomplish that we define the Bi-dimensional tabular coordinates, separate horizontal, vertical metadata and data contexts by introducing a new visibility matrix, encode units and nesting through the embeddings specifically optimized for mimicking intricacies of such complex structured data. Through evaluation on 5 large-scale structured datasets and 3 popular downstream tasks, we observed that our solution outperforms the state-of-the-art models with the significant MAP delta of up to 0.28. GPT-4 LLM+RAG slightly outperforms us with MRR delta of up to 0.1, while we outperform it with the MAP delta of up to 0.42.

Quantum well of AlGaAs/GaAs is very important to study transport properties of electrons due to its wider application in electronic devices. Hence, the double well of AlGaAs/GaAs with triple barrier is taken to study transmission probability. Transmission probability is found to decrease with the increase in the height and width of the barrier. Transmission probability with energy of electron shows two peaks while taking all three barrier of the same height. Whereas a single and higher value of peak is found when the height of the central barrier is slightly reduced.

Mathematical reasoning in Large Language Models (LLMs) is often evaluated using benchmarks with limited numerical ranges, failing to reflect real-world problem-solving across diverse scales. Furthermore, most existing evaluation methods only compare model outputs to ground-truth answers, obscuring insights into reasoning processes. To address these limitations, we introduce GSM-Ranges, a dataset generator derived from GSM8K that systematically perturbs numerical values in math problems to assess model robustness across varying numerical scales. Additionally, we propose a novel grading methodology that distinguishes between logical and non-logical errors, offering a more precise evaluation of reasoning processes beyond computational accuracy. Our experiments with various models reveal a significant increase in logical error rates-up to 14 percentage points-as numerical complexity rises, demonstrating a general weakness in reasoning with out-of-distribution numerical values. Moreover, while models demonstrate high accuracy on standalone arithmetic tasks, their performance deteriorates substantially when computations are embedded within word problems. These findings provide a comprehensive evaluation of LLMs' mathematical reasoning capabilities and inform future research directions for improving numerical generalization in language models.

This work focuses on generating realistic, physically-based human behaviors from multi-modal inputs, which may only partially specify the desired motion. For example, the input may come from a VR controller providing arm motion and body velocity, partial key-point animation, computer vision applied to videos, or even higher-level motion goals. This requires a versatile low-level humanoid controller that can handle such sparse, under-specified guidance, seamlessly switch between skills, and recover from failures. Current approaches for learning humanoid controllers from demonstration data capture some of these characteristics, but none achieve them all. To this end, we introduce the Masked Humanoid Controller (MHC), a novel approach that applies multi-objective imitation learning on augmented and selectively masked motion demonstrations. The training methodology results in an MHC that exhibits the key capabilities of catch-up to out-of-sync input commands, combining elements from multiple motion sequences, and completing unspecified parts of motions from sparse multimodal input. We demonstrate these key capabilities for an MHC learned over a dataset of 87 diverse skills and showcase different multi-modal use cases, including integration with planning frameworks to highlight MHC's ability to solve new user-defined tasks without any finetuning.

Linear recurrent neural networks enable powerful long-range sequence modeling with constant memory usage and time-per-token during inference. These architectures hold promise for streaming applications at the edge, but deployment in resource-constrained environments requires hardware-aware optimizations to minimize latency and energy consumption. Unstructured sparsity offers a compelling solution, enabling substantial reductions in compute and memory requirements--when accelerated by compatible hardware platforms. In this paper, we conduct a scaling study to investigate the Pareto front of performance and efficiency across inference compute budgets. We find that highly sparse linear RNNs consistently achieve better efficiency-performance trade-offs than dense baselines, with 2x less compute and 36% less memory at iso-accuracy. Our models achieve state-of-the-art results on a real-time streaming task for audio denoising. By quantizing our sparse models to fixed-point arithmetic and deploying them on the Intel Loihi 2 neuromorphic chip for real-time processing, we translate model compression into tangible gains of 42x lower latency and 149x lower energy consumption compared to a dense model on an edge GPU. Our findings showcase the transformative potential of unstructured sparsity, paving the way for highly efficient recurrent neural networks in real-world, resource-constrained environments., Comment: Under review

This paper introduces a mobile-based solution that enhances online shoe shopping through 3D modeling and Augmented Reality (AR), leveraging the efficiency of 3D Gaussian Splatting. Addressing the limitations of static 2D images, the framework generates realistic 3D shoe models from 2D images, achieving an average Peak Signal-to-Noise Ratio (PSNR) of 32, and enables immersive AR interactions via smartphones. A custom shoe segmentation dataset of 3120 images was created, with the best-performing segmentation model achieving an Intersection over Union (IoU) score of 0.95. This paper demonstrates the potential of 3D modeling and AR to revolutionize online shopping by offering realistic virtual interactions, with applicability across broader fashion categories.

Hydrogen production via the Hydrogen Evolution Reaction (HER) is critical for sustainable energy solutions, yet the reliance on expensive platinum (Pt) catalysts limits scalability. Zirconium-doped (\ce{Zr}-doped) MXenes, such as \ce{Ti3C2} and \ce{Ti3CN}, emerge as transformative alternatives, combining abundance, tunable electronic properties, and high catalytic potential. Using first-principles density functional theory (DFT), we show that \ce{Zr} doping at 3\% and 7\% significantly enhances HER activity by reducing the work function to the optimal range of 3.5-4.5~eV and achieving near-zero Gibbs free energy (\dgh) values of 0.18-0.16~eV, conditions ideal for efficient hydrogen adsorption and desorption. Bader charge analysis reveals substantial charge redistribution with enhanced electron accumulation at \ce{Zr} and \ce{N} sites, further driving catalytic performance. This synergy between optimized electronic structure and catalytic properties establishes \ce{Zr}-doped MXenes as cost-effective, high-performance alternatives to noble metals for HER. By combining exceptional catalytic efficiency with scalability, our work positions \ce{Zr}-doped MXenes as a breakthrough for green hydrogen production, offering a robust pathway toward renewable energy technologies and advancing the design of next-generation non-precious metal catalysts.

A heliospheric termination shock (HTS) surrounds our solar system at approximately 100 astronomical units from the Sun, where the expanding solar wind (SW) is compressed and heated before encountering the interstellar medium. HTS-accelerated particles govern the pressure balance with the interstellar medium, but little is known about the HTS's global properties beyond in situ measurements from Voyager in only two directions of the sky. We fill this gap in knowledge with a novel and complex methodology: particle-in-cell, test particle, and MHD simulations, combined with a global minimization scheme to derive global HTS compression ratio sky maps. The methods utilize Interstellar Boundary Explorer observations of energetic neutral atoms produced from HTS-accelerated particles. Our results reveal unique, three-dimensional characteristics, such as higher compression near the poles during solar minimum, north-south asymmetries from the disparate polar coronal holes' evolution, and minimum compression near the flanks likely from SW slowing by mass-loading over a greater distance to the HTS., Comment: 46 pages, 17 figures, 1 table, submitted to Nature Astronomy

Inclusive electron scattering cross sections off a hydrogen target at a beam energy of 10.6 GeV have been measured with data collected from the CLAS12 spectrometer at Jefferson Laboratory. These first absolute cross sections from CLAS12 cover a wide kinematic area in invariant mass W of the final state hadrons from the pion threshold up to 2.5 GeV for each bin in virtual photon four-momentum transfer squared $Q^2$ from 2.55 to 10.4~GeV$^2$ owing to the large scattering angle acceptance of the CLAS12 detector. Comparison of the cross sections with the resonant contributions computed from the CLAS results on the nucleon resonance electroexcitation amplitudes has demonstrated a promising opportunity to extend the information on their $Q^2$ evolution up to 10 GeV$^2$. Together these results from CLAS and CLAS12 offer good prospects for probing the nucleon parton distributions at large fractional parton momenta $x$ for $W$ < 2.5 GeV, while covering the range of distances where the transition from the strongly coupled to the perturbative regimes is expected.

Spectrum cartography (SC) focuses on estimating the radio power propagation map of multiple emitters across space and frequency using limited sensor measurements. Recent advances in SC have shown that leveraging learned deep generative models (DGMs) as structural constraints yields state-of-the-art performance. By harnessing the expressive power of neural networks, these structural "priors" capture intricate patterns in radio maps. However, training DGMs requires substantial data, which is not always available, and distribution shifts between training and testing data can further degrade performance. To address these challenges, this work proposes using untrained neural networks (UNNs) for SC. UNNs, commonly applied in vision tasks to represent complex data without training, encode structural information of data in neural architectures. In our approach, a custom-designed UNN represents radio maps under a spatio-spectral domain factorization model, leveraging physical characteristics to reduce sample complexity of SC. Experiments show that the method achieves performance comparable to learned DGM-based SC, without requiring training data., Comment: 6 pages 6 figures

The integration of Artificial Intelligence (AI) systems into technologies used by young digital citizens raises significant privacy concerns. This study investigates these concerns through a comparative analysis of stakeholder perspectives. A total of 252 participants were surveyed, with the analysis focusing on 110 valid responses from parents/educators and 100 from AI professionals after data cleaning. Quantitative methods, including descriptive statistics and Partial Least Squares Structural Equation Modeling, examined five validated constructs: Data Ownership and Control, Parental Data Sharing, Perceived Risks and Benefits, Transparency and Trust, and Education and Awareness. Results showed Education and Awareness significantly influenced data ownership and risk assessment, while Data Ownership and Control strongly impacted Transparency and Trust. Transparency and Trust, along with Perceived Risks and Benefits, showed minimal influence on Parental Data Sharing, suggesting other factors may play a larger role. The study underscores the need for user-centric privacy controls, tailored transparency strategies, and targeted educational initiatives. Incorporating diverse stakeholder perspectives offers actionable insights into ethical AI design and governance, balancing innovation with robust privacy protections to foster trust in a digital age., Comment: To appear in the 2025 IEEE 14th Annual Computing and Communication Workshop and Conference (CCWC) proceedings

As Artificial Intelligence (AI) systems become increasingly integrated into various aspects of daily life, concerns about privacy and ethical accountability are gaining prominence. This study explores stakeholder perspectives on privacy in AI systems, focusing on educators, parents, and AI professionals. Using qualitative analysis of survey responses from 227 participants, the research identifies key privacy risks, including data breaches, ethical misuse, and excessive data collection, alongside perceived benefits such as personalized services, enhanced efficiency, and educational advancements. Stakeholders emphasized the need for transparency, privacy-by-design, user empowerment, and ethical oversight to address privacy concerns effectively. The findings provide actionable insights into balancing the benefits of AI with robust privacy protections, catering to the diverse needs of stakeholders. Recommendations include implementing selective data use, fostering transparency, promoting user autonomy, and integrating ethical principles into AI development. This study contributes to the ongoing discourse on ethical AI, offering guidance for designing privacy-centric systems that align with societal values and build trust among users. By addressing privacy challenges, this research underscores the importance of developing AI technologies that are not only innovative but also ethically sound and responsive to the concerns of all stakeholders., Comment: To appear in the 2025 IEEE 14th Annual Computing and Communication Workshop and Conference (CCWC) proceedings

Large language models (LLMs) are increasingly prevalent in recommender systems, where LLMs can be used to generate personalized recommendations. Here, we examine how different LLM-generated explanations for movie recommendations affect users' perceptions of cognitive, affective, and utilitarian needs and consumption intentions. In a pre-registered, between-subject online experiment (N=759) and follow-up interviews (N=30), we compare (a) LLM-generated generic explanations, and (b) LLM-generated contextualized explanations. Our findings show that contextualized explanations (i.e., explanations that incorporate users' past behaviors) effectively meet users' cognitive needs while increasing users' intentions to watch recommended movies. However, adding explanations offers limited benefits in meeting users' utilitarian and affective needs, raising concerns about the proper design and implications of LLM-generated explanations. Qualitative insights from interviews reveal that referencing users' past preferences enhances trust and understanding but can feel excessive if overused. Furthermore, users with more active and positive engagement with the recommender system and movie-watching get substantial gains from contextualized explanations. Overall, our research clarifies how LLM-generated recommendations influence users' motivations and behaviors, providing valuable insights for the future development of user-centric recommender systems, a key element in social media platforms and online ecosystems.

The ocean is teeming with a myriad of invertebrate planktonic larvae, which thrive in a viscous fluid environment. Many of them rely on ciliary beating to generate fluid flows for locomotion and feeding. Their forms, local morphologies, and ciliation patterns exhibit remarkable diversity, producing intricate and dynamic three-dimensional flows that are notoriously difficult to characterize in laboratory settings. Traditional microscopic imaging techniques typically involve gently squeeze-confining the soft larvae between a glass slide and cover slip, to study their flows in quasi-two-dimensions. However, a comprehensive hydrodynamic framework for the low-to-intermediate Reynolds number flows at the larval scale in quasi-2D confinement - particularly in light of their complex forms - has remained elusive. Here, we demonstrate that vortices around larvae proliferate with increasing confinement, and elucidate the underlying physical mechanism. We experimentally quantify confinement-induced flows in larvae of sea stars and sea urchins. The flows exhibited strikingly universal patterns: under weak confinement, all larvae generated two vortices, whereas under strong confinement, multiple vortices appeared. The experimental observations were well captured by a low Reynolds number theoretical model based on superposition of confined Stokeslets. Building on experiments and theory, we developed a comprehensive framework for confinement-induced larval flows, which suggests that the vorticity dynamics are primarily determined by local morphological features, rather than the larval forms. Our work provides fundamental insights into the form-functional relationships between larval morphology and flow generation. Our findings are broadly applicable to understanding fluid flows generated by a wide range of ciliated organisms with complex forms and morphologies, from micro- to milli-length-scales., Comment: 26 pages, 3 main figures, 5 extended data figures. (First two authors contributed equally)

The rapid spread of fake news presents a significant global challenge, particularly in low-resource languages like Bangla, which lack adequate datasets and detection tools. Although manual fact-checking is accurate, it is expensive and slow to prevent the dissemination of fake news. Addressing this gap, we introduce BanFakeNews-2.0, a robust dataset to enhance Bangla fake news detection. This version includes 11,700 additional, meticulously curated fake news articles validated from credible sources, creating a proportional dataset of 47,000 authentic and 13,000 fake news items across 13 categories. In addition, we created a manually curated independent test set of 460 fake and 540 authentic news items for rigorous evaluation. We invest efforts in collecting fake news from credible sources and manually verified while preserving the linguistic richness. We develop a benchmark system utilizing transformer-based architectures, including fine-tuned Bidirectional Encoder Representations from Transformers variants (F1-87\%) and Large Language Models with Quantized Low-Rank Approximation (F1-89\%), that significantly outperforms traditional methods. BanFakeNews-2.0 offers a valuable resource to advance research and application in fake news detection for low-resourced languages. We publicly release our dataset and model on Github to foster research in this direction.

The impact of Artificial Intelligence (AI) is transforming various aspects of urban life, including, governance, policy and planning, healthcare, sustainability, economics, entrepreneurship, etc. Although AI immense potential for positively impacting urban living, its success depends on overcoming significant challenges, particularly in telecommunications infrastructure. Smart city applications, such as, federated learning, Internet of Things (IoT), and online financial services, require reliable Quality of Service (QoS) from telecommunications networks to ensure effective information transfer. However, with over three billion people underserved or lacking access to internet, many of these AI-driven applications are at risk of either remaining underutilized or failing altogether. Furthermore, many IoT and video-based applications in densely populated urban areas require high-quality connectivity. This paper explores these issues, focusing on the challenges that need to be mitigated to make AI succeed in emerging countries, where more than 80% of the world population resides and urban migration grows. In this context, an overview of a case study conducted in Kathmandu, Nepal, highlights citizens' aspirations for affordable, high-quality internet-based services. The findings underscore the pressing need for advanced telecommunication networks to meet diverse user requirements while addressing investment and infrastructure gaps. This discussion provides insights into bridging the digital divide and enabling AI's transformative potential in urban areas.

The integration of distributed energy resources (DERs), particularly photovoltaic (PV) systems, into power grids has gained major attention due to their environmental and economic benefits. Although traditional transformer-based grid-connected PV inverters provide galvanic isolation for leakage current, they suffer from major drawbacks of high cost, lower efficiency, and increased size. Transformer-less grid-connected PV inverters (TLGI) have emerged as a prominent alternative, as they achieve higher efficiency, compact design, and lower cost. However, due to a lack of galvanic isolation, TLGIs are highly affected by leakage current caused by the fluctuation of common-mode voltage (CMV). This paper investigates three topologies H4, H5, and HERIC with comparisons between their CMV, differential-mode voltage (DMV), total harmonic distortion (THD), and leakage current. A simulation was conducted for each topology in MATLAB/Simulink R2023a, and the results demonstrate that the H5 topology achieves a balance between low leakage current, reduced THD, and optimal operational efficiency, making it suitable for practical application., Comment: 17 pages

Science laboratory automation enables accelerated discovery in life sciences and materials. However, it requires interdisciplinary collaboration to address challenges such as robust and flexible autonomy, reproducibility, throughput, standardization, the role of human scientists, and ethics. This article highlights these issues, reflecting perspectives from leading experts in laboratory automation across different disciplines of the natural sciences.

We propose a measurement of exclusive electroproduction of $\phi$ mesons near threshold in Hall C. The $|t|$-dependence of the exclusive $\phi$ cross section, $d\sigma/d|t|$, has recently been proposed as an observable sensitive to the strangeness $D$-term. The contribution of strangeness to the total quark $D$-term is presently unknown, with different arguments favoring $D_s$ being large, small, or even having opposite sign from the total quark $D$-term. In addition, this dataset will allow us to perform measurements of other exclusive meson final states, including the first measurement of $\eta'$ electroproduction., Comment: Letter-of-intent submitted to Jefferson Lab PAC 52

We focus on enhancing comprehension in small-group recorded conversations, which serve as a medium to bring people together and provide a space for sharing personal stories and experiences on crucial social matters. One way to parse and convey information from these conversations is by sharing highlighted excerpts in subsequent conversations. This can help promote a collective understanding of relevant issues, by highlighting perspectives and experiences to other groups of people who might otherwise be unfamiliar with and thus unable to relate to these experiences. The primary challenge that arises then is that excerpts taken from one conversation and shared in another setting might be missing crucial context or key elements that were previously introduced in the original conversation. This problem is exacerbated when conversations become lengthier and richer in themes and shared experiences. To address this, we explore how Large Language Models (LLMs) can enrich these excerpts by providing socially relevant context. We present approaches for effective contextualization to improve comprehension, readability, and empathy. We show significant improvements in understanding, as assessed through subjective and objective evaluations. While LLMs can offer valuable context, they struggle with capturing key social aspects. We release the Human-annotated Salient Excerpts (HSE) dataset to support future work. Additionally, we show how context-enriched excerpts can provide more focused and comprehensive conversation summaries., Comment: Accepted at COLING 2025

The new transmission control protocol (TCP) relies on Deep Learning (DL) for prediction and optimization, but requires significant manual effort to design deep neural networks (DNNs) and struggles with generalization in dynamic environments. Inspired by the success of large language models (LLMs), this study proposes TCP-LLM, a novel framework leveraging LLMs for TCP applications. TCP-LLM utilizes pre-trained knowledge to reduce engineering effort, enhance generalization, and deliver superior performance across diverse TCP tasks. Applied to reducing flow unfairness, adapting congestion control, and preventing starvation, TCP-LLM demonstrates significant improvements over TCP with minimal fine-tuning.

This systematic literature review investigates perceptions, concerns, and expectations of young digital citizens regarding privacy in artificial intelligence (AI) systems, focusing on social media platforms, educational technology, gaming systems, and recommendation algorithms. Using a rigorous methodology, the review started with 2,000 papers, narrowed down to 552 after initial screening, and finally refined to 108 for detailed analysis. Data extraction focused on privacy concerns, data-sharing practices, the balance between privacy and utility, trust factors in AI, transparency expectations, and strategies to enhance user control over personal data. Findings reveal significant privacy concerns among young users, including a perceived lack of control over personal information, potential misuse of data by AI, and fears of data breaches and unauthorized access. These issues are worsened by unclear data collection practices and insufficient transparency in AI applications. The intention to share data is closely associated with perceived benefits and data protection assurances. The study also highlights the role of parental mediation and the need for comprehensive education on data privacy. Balancing privacy and utility in AI applications is crucial, as young digital citizens value personalized services but remain wary of privacy risks. Trust in AI is significantly influenced by transparency, reliability, predictable behavior, and clear communication about data usage. Strategies to improve user control over personal data include access to and correction of data, clear consent mechanisms, and robust data protection assurances. The review identifies research gaps and suggests future directions, such as longitudinal studies, multicultural comparisons, and the development of ethical AI frameworks., Comment: To appear in the 2024 IEEE Annual Information Technology, Electronics and Mobile Communication Conference proceedings

We develop a hydrodynamic description of self-generated electrolyte flow in capillaries whose bounding walls feature both non-uniform distributions of charge and non-uniform active ionic fluxes. The hydrodynamic velocity arising in such a system has components that are forbidden by symmetry in the absence of charge and fluxes. We show that flow states with unidirectional and circulatory components emerge when modulated boundary patterns are offset by a selection of flux-charge phase differences associated with the spatial mismatch between sites of peak flux and charge densities on the wall. Self-induced local ionic elevation and depletion, constantly disrupting a non-uniform double layer, promotes directed gradients yielding persistent body forces that generate longitudinal fluid motion along the capillary. Our work quantifies a boundary-driven mechanism based on active-charged patterns that produce self-sustained electrolyte flow in confined environments and exists even in the absence of any external bulk-imposed fields or gradients. It provides a theoretical framework for understanding the effect of ionic boundary properties that are relevant in biological or soft matter systems and can be utilized in nanofluidics and iontronics.

Progress in AI is driven largely by the scale and quality of training data. Despite this, there is a deficit of empirical analysis examining the attributes of well-established datasets beyond text. In this work we conduct the largest and first-of-its-kind longitudinal audit across modalities--popular text, speech, and video datasets--from their detailed sourcing trends and use restrictions to their geographical and linguistic representation. Our manual analysis covers nearly 4000 public datasets between 1990-2024, spanning 608 languages, 798 sources, 659 organizations, and 67 countries. We find that multimodal machine learning applications have overwhelmingly turned to web-crawled, synthetic, and social media platforms, such as YouTube, for their training sets, eclipsing all other sources since 2019. Secondly, tracing the chain of dataset derivations we find that while less than 33% of datasets are restrictively licensed, over 80% of the source content in widely-used text, speech, and video datasets, carry non-commercial restrictions. Finally, counter to the rising number of languages and geographies represented in public AI training datasets, our audit demonstrates measures of relative geographical and multilingual representation have failed to significantly improve their coverage since 2013. We believe the breadth of our audit enables us to empirically examine trends in data sourcing, restrictions, and Western-centricity at an ecosystem-level, and that visibility into these questions are essential to progress in responsible AI. As a contribution to ongoing improvements in dataset transparency and responsible use, we release our entire multimodal audit, allowing practitioners to trace data provenance across text, speech, and video., Comment: ICLR 2025. 10 pages, 5 figures (main paper)

Blockchain technology has emerged as a transformative tool for data management in a variety of industries, including fintech, research and healthcare. We have developed a workable blockchain based system that utilizes non fungible tokens NFTs to tokenize and prove ownership of the academic institutions credentials. This makes it easier to create provenance and ownership documentation for academic data and meta credentials. This system enables the secure sharing of academic information while maintaining control, offering incentives for collaboration, and granting users full transparency and control over data access. While the initial adoption of these systems is crucial for ongoing service usage, the exploration of the user acceptance behavioural model remains limited in the existing literature. In this paper, we build upon the Technology Acceptance Model TAM, incorporating additional elements to scrutinize the impact of perceived ease of use, perceived usability, and attitude towards the system on the intention to use a blockchain based academic data and meta credentials sharing system. The research, grounded in user evaluations of a prototype, employs a TAM validated questionnaire. Results indicate that individual constructs notably affect the intention to use the system, and their collective impact is statistically significant. Specifically, perceived ease of use is the sole factor with an insignificant influence on the intention to use. The paper underscores the dominant influence of attitude towards the system on perceived usefulness. It concludes with a discussion on the implications of these findings within the context of blockchain based academic data and meta credentials sharing, incorporating NFTs for ownership definition., Comment: Blockchain and Applications 6th International Congress Conference

Ultrasound imaging of the medial elbow is crucial for the early diagnosis of Ulnar Collateral Ligament (UCL) injuries. Specifically, measuring the elbow joint space in ultrasound images is used to assess the valgus instability of the elbow caused by UCL injuries. To automate this measurement, a model trained on a precisely annotated dataset is necessary; however, no publicly available dataset exists to date. This study introduces a novel ultrasound medial elbow dataset to measure the joint space. The dataset comprises 4,201 medial elbow ultrasound images from 22 subjects, with landmark annotations on the humerus and ulna, based on the expertise of three orthopedic surgeons. We evaluated joint space measurement methods on our proposed dataset using heatmap-based, regression-based, and token-based landmark detection methods. While heatmap-based landmark detection methods generally achieve high accuracy, they sometimes produce multiple peaks on a heatmap, leading to incorrect detection. To mitigate this issue and enhance landmark localization, we propose Shape Subspace (SS) landmark refinement by measuring geometrical similarities between the detected and reference landmark positions. The results show that the mean joint space measurement error is 0.116 mm when using HRNet. Furthermore, SS landmark refinement can reduce the mean absolute error of landmark positions by 0.010 mm with HRNet and by 0.103 mm with ViTPose on average. These highlight the potential for high-precision, real-time diagnosis of UCL injuries by accurately measuring joint space. Lastly, we demonstrate point-based segmentation for the humerus and ulna using the detected landmarks as inputs. Our dataset will be publicly available at https://github.com/Akahori000/Ultrasound-Medial-Elbow-Dataset

Altermagnets are a recently discovered class of magnetic material with great potential for applications in the field of spintronics, owing to their non-relativistic spin-splitting and simultaneous antiferromagnetic order. One of the most studied candidates for altermagnetic materials is rutile structured RuO2. However, it has recently come under significant scrutiny as evidence emerged for its lack of any magnetic order. In this work, we study bilayers of epitaxial RuO2 and ferromagnetic permalloy (Fe19Ni81) by time-domain terahertz spectroscopy, probing for three possible mechanisms of laser-induced charge dynamics: the inverse spin Hall effect (ISHE), electrical anisotropic conductivity (EAC), and inverse altermagnetic spin-splitting effect (IASSE). We examine films of four common RuO2 layer orientations: (001), (100), (110), and (101). If RuO2 is altermagnetic, then the (100) and (101) oriented samples are expected to produce anisotropic emission from the IASSE, however, our results do not indicate the presence of IASSE for either as-deposited or field annealed samples. The THz emission from all samples is instead consistent with charge dynamics induced by only the relativistic ISHE and the non-relativistic and non-magnetic EAC, casting further doubt on the existence of altermagnetism in RuO2. In addition, we find that in the (101) oriented RuO2 sample, the combination of ISHE and EAC emission mechanisms produces THz emission which is tunable between linear and elliptical polarization by modulation of the external magnetic field.

Creation and curation of knowledge graphs can accelerate disease discovery and analysis in real-world data. While disease ontologies aid in biological data annotation, codified categories (SNOMED-CT, ICD10, CPT) may not capture patient condition nuances or rare diseases. Multiple disease definitions across data sources complicate ontology mapping and disease clustering. We propose creating patient knowledge graphs using large language model extraction techniques, allowing data extraction via natural language rather than rigid ontological hierarchies. Our method maps to existing ontologies (MeSH, SNOMED-CT, RxNORM, HPO) to ground extracted entities. Using a large ambulatory care EHR database with 33.6M patients, we demonstrate our method through the patient search for Dravet syndrome, which received ICD10 recognition in October 2020. We describe our construction of patient-specific knowledge graphs and symptom-based patient searches. Using confirmed Dravet syndrome ICD10 codes as ground truth, we employ LLM-based entity extraction to characterize patients in grounded ontologies. We then apply this method to identify Beta-propeller protein-associated neurodegeneration (BPAN) patients, demonstrating real-world discovery where no ground truth exists., Comment: 10 pages, 3 figures, Work published in 4th Workshop on Knowledge Graphs and Big Data (In Conjunction with IEEE Big Data 2024)

While the subclass of interacting supernovae with narrow hydrogen emission lines (SNe IIn) consists of some of the longest-lasting and brightest SNe ever discovered, their progenitors are still not well understood. Investigating SNe IIn as they emit across the electromagnetic spectrum is the most robust way to understand the progenitor evolution before the explosion. This work presents X-Ray, optical, infrared, and radio observations of the strongly interacting Type IIn SN 2020ywx covering a period $>1200$ days after discovery. Through multiwavelength modeling, we find that the progenitor of 2020ywx was losing mass at $\sim10^{-2}$--$10^{-3} \mathrm{\,M_{\odot}\,yr^{-1}}$ for at least 100 yr pre-explosion using the circumstellar medium (CSM) speed of $120$ km/s measured from our optical and NIR spectra. Despite the similar magnitude of mass-loss measured in different wavelength ranges, we find discrepancies between the X-ray and optical/radio-derived mass-loss evolution, which suggest asymmetries in the CSM. Furthermore, we find evidence for dust formation due to the combination of a growing blueshift in optical emission lines and near-infrared continuum emission which we fit with blackbodies at $\sim$ 1000 K. Based on the observed elevated mass loss over more than 100 years and the configuration of the CSM inferred from the multiwavelength observations, we invoke binary interaction as the most plausible mechanism to explain the overall mass-loss evolution. SN 2020ywx is thus a case that may support the growing observational consensus that SNe IIn mass loss is explained by binary interaction., Comment: Submitted to ApJ, 31 pages, 19 figures

The EAGLE-1 mission aims to develop Europe's first sovereign, end-to-end space-based quantum key distribution (QKD) system. The mission is led by the European Space Agency (ESA) and SES in collaboration with several European National Space Agencies and private partners. The state-of-the-art QKD system will consist of a payload on board the EAGLE-1 low Earth orbit (LEO) satellite, optical ground stations, quantum operational networks, and key management system. The EAGLE-1 mission represents a major step for next-generation quantum communication infrastructures, delivering valuable technical results and mission data. The Institute of Communications and Navigation (IKN) of the German Aerospace Center (DLR) is a key partner in the EAGLE-1 mission and is involved in the research and development of elements in both space and ground segments. Here we report on the development of the QKD transmitter, a vital part of the QKD payload, and the customization of the Optical Ground Station Oberpfaffenhofen (OGS-OP) to conduct the IOT phase of EAGLE-1. For the space segment, DLR-IKN is in charge of the design of the QKD transmitter, including the development of the software and firmware. This transmitter generates quantum states which are used to implement a QKD protocol based on an optical signal, that will be transmitted to ground. For the ground segment, The OGS-OP will serve as the in-orbit testing ground station for EAGLE-1. Building upon the expertise with a range of satellites for quantum communication, as well as new implementations, OGS-OP will validate the performance of the payload, optical link and QKD system for the first time. We present the main developments of OGS-OP for the mission, which includes the implementation of an upgraded adaptive optics system to correct for atmospheric distortions and optimize the coupling of the incoming light into a single mode optical fiber., Comment: 9 pages, 5 figures. Manuscript presented at the 75th International Astronautical Congress (IAC), Milan, Italy, 14-18 October 2024

We solve a system of massless fermions constrained to two space-time dimensions interacting via a $d$ space-time dimensional Maxwell field. Through dimensional reduction to the defect and bosonization, the system maps to a massless scalar interacting with a nonlocal Maxwell field through a $F \phi$-coupling. The $d=2$ dimensional case is the usual Schwinger model where the photon gets a mass. More generally, in $2