Search

Latest developments in Large Multimodal Models (LMMs) have broadened their capabilities to include video understanding. Specifically, Text-to-video (T2V) models have made significant progress in quality, comprehension, and duration, excelling at creating videos from simple textual prompts. Yet, they still frequently produce hallucinated content that clearly signals the video is AI-generated. We introduce ViBe: a large-scale Text-to-Video Benchmark of hallucinated videos from T2V models. We identify five major types of hallucination: Vanishing Subject, Numeric Variability, Temporal Dysmorphia, Omission Error, and Physical Incongruity. Using 10 open-source T2V models, we developed the first large-scale dataset of hallucinated videos, comprising 3,782 videos annotated by humans into these five categories. ViBe offers a unique resource for evaluating the reliability of T2V models and provides a foundation for improving hallucination detection and mitigation in video generation. We establish classification as a baseline and present various ensemble classifier configurations, with the TimeSFormer + CNN combination yielding the best performance, achieving 0.345 accuracy and 0.342 F1 score. This benchmark aims to drive the development of robust T2V models that produce videos more accurately aligned with input prompts.

Device driver bugs are the leading cause of OS compromises, and their formal verification is therefore highly desirable. To the best of our knowledge, no realistic and performant driver has been verified for a non-trivial device. We propose Pancake, an imperative language for systems programming that features a well-defined and verification-friendly semantics. Leveraging the verified compiler backend of the CakeML functional language, we develop a compiler for Pancake that guarantees that the binary retains the semantics of the source code. Usng automatic translation of Pancake to the Viper SMT front-end, we verify a performant driver for an Ethernet NIC., Comment: 15 pages, 5 figures

Rapid advancements in large language models have unlocked remarkable capabilities when it comes to processing and summarizing unstructured text data. This has implications for the analysis of rich, open-ended datasets, such as survey responses, where LLMs hold the promise of efficiently distilling key themes and sentiments. However, as organizations increasingly turn to these powerful AI systems to make sense of textual feedback, a critical question arises, can we trust LLMs to accurately represent the perspectives contained within these text based datasets? While LLMs excel at generating human-like summaries, there is a risk that their outputs may inadvertently diverge from the true substance of the original responses. Discrepancies between the LLM-generated outputs and the actual themes present in the data could lead to flawed decision-making, with far-reaching consequences for organizations. This research investigates the effectiveness of LLMs as judge models to evaluate the thematic alignment of summaries generated by other LLMs. We utilized an Anthropic Claude model to generate thematic summaries from open-ended survey responses, with Amazon's Titan Express, Nova Pro, and Meta's Llama serving as LLM judges. The LLM-as-judge approach was compared to human evaluations using Cohen's kappa, Spearman's rho, and Krippendorff's alpha, validating a scalable alternative to traditional human centric evaluation methods. Our findings reveal that while LLMs as judges offer a scalable solution comparable to human raters, humans may still excel at detecting subtle, context-specific nuances. This research contributes to the growing body of knowledge on AI assisted text analysis. We discuss limitations and provide recommendations for future research, emphasizing the need for careful consideration when generalizing LLM judge models across various contexts and use cases., Comment: 11 pages, 1 appendix

The ErbB receptor family, including EGFR and HER2, plays a crucial role in cell growth and survival and is associated with the progression of various cancers such as breast and lung cancer. In this study, we developed a deep learning model to predict the binding affinity of ErbB inhibitors using molecular fingerprints derived from SMILES representations. The SMILES representations for each ErbB inhibitor were obtained from the ChEMBL database. We first generated Morgan fingerprints from the SMILES strings and applied AutoDock Vina docking to calculate the binding affinity values. After filtering the dataset based on binding affinity, we trained a deep neural network (DNN) model to predict binding affinity values from the molecular fingerprints. The model achieved significant performance, with a Mean Squared Error (MSE) of 0.2591, Mean Absolute Error (MAE) of 0.3658, and an R-squared value of 0.9389 on the training set. Although performance decreased slightly on the test set (R squared = 0.7731), the model still demonstrated robust generalization capabilities. These results indicate that the deep learning approach is highly effective for predicting the binding affinity of ErbB inhibitors, offering a valuable tool for virtual screening and drug discovery., Comment: 40 pages, 1 graphical abstract, 2 figures, 6 appendices

The Ethereum blockchain network enables transaction processing and smart-contract execution through levies of transaction fees, commonly known as gas fees. This framework mediates economic participation via a market-based mechanism for gas fees, permitting users to offer higher gas fees to expedite pro-cessing. Historically, the ensuing gas fee volatility led to critical disequilibria between supply and demand for block space, presenting stakeholder challenges. This study examines the dynamic causal interplay between transaction fees and economic subsystems leveraging the network. By utilizing data related to unique active wallets and transaction volume of each subsystem and applying time-varying Granger causality analysis, we reveal temporal heterogeneity in causal relationships between economic activity and transaction fees across all subsystems. This includes (a) a bidirectional causal feedback loop between cross-blockchain bridge user activity and transaction fees, which diminishes over time, potentially signaling user migration; (b) a bidirectional relationship between centralized cryptocurrency exchange deposit and withdrawal transaction volume and fees, indicative of increased competition for block space; (c) decentralized exchange volumes causally influence fees, while fees causally influence user activity, although this relationship is weakening, potentially due to the diminished significance of decentralized finance; (d) intermittent causal relationships with maximal extractable value bots; (e) fees causally in-fluence non-fungible token transaction volumes; and (f) a highly significant and growing causal influence of transaction fees on stablecoin activity and transaction volumes highlight its prominence., Comment: 52 pages, 4 tables, 9 figures

Tether Limited has the sole authority to create (mint) and destroy (burn) Tether stablecoins (USDT). This paper investigates Bitcoin's response to USDT supply change events between 2014 and 2021 and identifies an interesting asymmetry between Bitcoin's responses to USDT minting and burning events. Bitcoin responds positively to USDT minting events over 5- to 30-minute event windows, but this response begins declining after 60 minutes. State-dependence is also demonstrated, with Bitcoin prices exhibiting a greater increase when the corresponding USDT minting event coincides with positive investor sentiment and is announced to the public by data service provider, Whale Alert, on Twitter., Comment: 18 pages, 5 tables

Several astrophysical and cosmological observations suggest the existence of dark matter (DM) through its gravitational effects, yet its nature remains elusive. Despite the lack of DM signals from direct detection experiments, efforts continue to focus on the indirect detection of DM from DM-rich astrophysical objects. Dwarf spheroidal galaxies (dSphs) are among the most promising targets for such searches. In this work, we aim to investigate the expected DM capture rate from the stellar component of 10 nearby DM-rich dSphs, assuming that the accumulated DM eventually annihilates into light, long-lived mediators (LLLMs) which decay into gamma rays outside the dSphs. We analyze nearly 16 years of {\it Fermi}-LAT data to search for DM annihilation through LLLMs, and, from the observed stacked flux upper limits, set limits on the DM-nucleon scattering cross section using model-independent methods. Additionally, we incorporate the Sommerfeld Enhancement (SE) effect into the DM annihilation process, obtaining bounds on the DM-nucleon scattering cross-section of $\sim~10^{-36} {\rm cm}^2$ for DM masses around 100 GeV. This allows us to explore an alternative avenue for exploring DM phenomena from dSphs and compare our results with the bounds reported by direct DM detection experiments and other celestial bodies., Comment: 25 pages, 6 figures, Comments are welcome

We present deep observations in targeted regions of the string landscape through a combination of analytic and dedicated numerical methods. Specifically, we devise an algorithm designed for the systematic construction of Type IIB flux vacua in finite regions of moduli space. Our algorithm is universally applicable across Calabi-Yau orientifold compactifications and can be used to enumerate flux vacua in a region given sufficient computational efforts. As a concrete example, we apply our methods to a two-modulus Calabi-Yau threefold, demonstrating that systematic enumeration is feasible and revealing intricate structures in vacuum distributions. Our results highlight local deviations from statistical expectations, providing insights into vacuum densities, superpotential distributions, and moduli mass hierarchies. This approach opens pathways for precise, data-driven mappings of the string landscape, complementing analytic studies and advancing the understanding of the distribution of flux vacua. This allows us to obtain different types of solutions with hierarchical suppressions, e.g.~vacua with small values of the Gukov-Vafa-Witten superpotential $|W_0|$. We find an example with $|W_0| = 5.547 \times 10^{-5}$ at large complex structure, without light directions and the use of non-perturbative effects., Comment: 26 pages, 10 figures, 4 tables

Reinforcement Learning with Human Feedback (RLHF) and its variants have made huge strides toward the effective alignment of large language models (LLMs) to follow instructions and reflect human values. More recently, Direct Alignment Algorithms (DAAs) have emerged in which the reward modeling stage of RLHF is skipped by characterizing the reward directly as a function of the policy being learned. Examples include Direct Preference Optimization (DPO) and Simple Preference Optimization (SimPO). These methods often suffer from likelihood displacement, a phenomenon by which the probabilities of preferred responses are often reduced undesirably. In this paper, we argue that, for DAAs the reward (function) shape matters. We introduce AlphaPO, a new DAA method that leverages an $\alpha$-parameter to help change the shape of the reward function beyond the standard log reward. AlphaPO helps maintain fine-grained control over likelihood displacement and over-optimization. Compared to SimPO, one of the best performing DAAs, AlphaPO leads to about 7\% to 10\% relative improvement in alignment performance for the instruct versions of Mistral-7B and Llama3-8B. The analysis and results presented highlight the importance of the reward shape, and how one can systematically change it to affect training dynamics, as well as improve alignment performance., Comment: Preprint. Work in progress

We review some of the results obtained to date with the aid of the PR-DNS approach to turbulent particulate flows. It is shown that the method has matured to a point which allows to apply it successfully to a wide variety of fluid/particle configurations, albeit still at a relatively large computational cost. Due to the availability of high-fidelity space-and-time-resolved data, a number of challenging open questions have already been addressed in unprecedented detail.

The rapid rise of large language models (LLMs) has unlocked many applications but also underscores the challenge of aligning them with diverse values and preferences. Direct Preference Optimization (DPO) is central to alignment but constrained by fixed divergences and limited feature transformations. We propose DPO-Kernels, which integrates kernel methods to address these issues through four key contributions: (i) Kernelized Representations with polynomial, RBF, Mahalanobis, and spectral kernels for richer transformations, plus a hybrid loss combining embedding-based and probability-based objectives; (ii) Divergence Alternatives (Jensen-Shannon, Hellinger, Renyi, Bhattacharyya, Wasserstein, and f-divergences) for greater stability; (iii) Data-Driven Selection metrics that automatically choose the best kernel-divergence pair; and (iv) a Hierarchical Mixture of Kernels for both local precision and global modeling. Evaluations on 12 datasets demonstrate state-of-the-art performance in factuality, safety, reasoning, and instruction following. Grounded in Heavy-Tailed Self-Regularization, DPO-Kernels maintains robust generalization for LLMs, offering a comprehensive resource for further alignment research.

Following an analysis of existing AI-related exchange-traded funds (ETFs), we reveal the selection criteria for determining which stocks qualify as AI-related are often opaque and rely on vague phrases and subjective judgments. This paper proposes a new, objective, data-driven approach using natural language processing (NLP) techniques to classify AI stocks by analyzing annual 10-K filings from 3,395 NASDAQ-listed firms between 2011 and 2023. This analysis quantifies each company's engagement with AI through binary indicators and weighted AI scores based on the frequency and context of AI-related terms. Using these metrics, we construct four AI stock indices-the Equally Weighted AI Index (AII), the Size-Weighted AI Index (SAII), and two Time-Discounted AI Indices (TAII05 and TAII5X)-offering different perspectives on AI investment. We validate our methodology through an event study on the launch of OpenAI's ChatGPT, demonstrating that companies with higher AI engagement saw significantly greater positive abnormal returns, with analyses supporting the predictive power of our AI measures. Our indices perform on par with or surpass 14 existing AI-themed ETFs and the Nasdaq Composite Index in risk-return profiles, market responsiveness, and overall performance, achieving higher average daily returns and risk-adjusted metrics without increased volatility. These results suggest our NLP-based approach offers a reliable, market-responsive, and cost-effective alternative to existing AI-related ETF products. Our innovative methodology can also guide investors, asset managers, and policymakers in using corporate data to construct other thematic portfolios, contributing to a more transparent, data-driven, and competitive approach., Comment: 43 pages, 5 tables, 3 figures, 1 appendix figure

We present OLMo 2, the next generation of our fully open language models. OLMo 2 includes dense autoregressive models with improved architecture and training recipe, pretraining data mixtures, and instruction tuning recipes. Our modified model architecture and training recipe achieve both better training stability and improved per-token efficiency. Our updated pretraining data mixture introduces a new, specialized data mix called Dolmino Mix 1124, which significantly improves model capabilities across many downstream task benchmarks when introduced via late-stage curriculum training (i.e. specialized data during the annealing phase of pretraining). Finally, we incorporate best practices from T\"ulu 3 to develop OLMo 2-Instruct, focusing on permissive data and extending our final-stage reinforcement learning with verifiable rewards (RLVR). Our OLMo 2 base models sit at the Pareto frontier of performance to compute, often matching or outperforming open-weight only models like Llama 3.1 and Qwen 2.5 while using fewer FLOPs and with fully transparent training data, code, and recipe. Our fully open OLMo 2-Instruct models are competitive with or surpassing open-weight only models of comparable size, including Qwen 2.5, Llama 3.1 and Gemma 2. We release all OLMo 2 artifacts openly -- models at 7B and 13B scales, both pretrained and post-trained, including their full training data, training code and recipes, training logs and thousands of intermediate checkpoints. The final instruction model is available on the Ai2 Playground as a free research demo., Comment: Model demo available at playground.allenai.org

This paper explores the prevalence of size-dependent aspects in the context of dust clusters with the help of Molecular Dynamics (MD) simulations in two dimensions. The transition from macroscale (identified by the dominance of the number of dust particles in bulk) to microscale (where the number of particles on the surface dominates) is explored systematically. The dust particles organize in a multi-ringed structure under transverse confinement. The ring size and the number of rings increase with increasing number of dust particles. Interestingly, the formation of an additional ring is always preceded by structures with a pentagonal symmetry in the core. A detailed study of this formation has been investigated under various symmetries of the boundary condition and different values of the shielding potential., Comment: 10 pages, 10 Figures, 3 tables, 48 References

Oxyhydrides are gaining attention for their diverse functionality. However, evaluating electron, ion, and phonon transport properties and comprehending their mechanism remains challenging due to the difficulty in oxyhydride synthesis involving the trade-off between achieving density sintering and preventing hydrogen evolution. High-temperature treatments, often required to reduce intergranular resistance, typically cause hydrogen evolution before sufficient sintering. Instead of a post-annealing process aiming for strong chemical connections between oxyhydride particles, this research demonstrates a novel synthesis technique that converts pre-sintered bulk oxides into oxyhydrides while preserving the dense state. The process employs a high-pressure diffusion control method, facilitating the co-diffusion of hydride and oxide ions., Comment: 4 Figures, 8 Figures in SI, 3 Tables in SI

This study aims to develop a deep learning model for predicting the binding affinity of ligands targeting the Peroxisome Proliferator-Activated Receptor (PPAR) family, using 2D molecular descriptors. A dataset of 3,764 small molecules with known binding affinities, sourced from the ChEMBL database, was preprocessed by eliminating duplicates and incomplete data. Molecular docking simulations using AutoDock Vina were performed to predict binding affinities for the PPAR receptor family. 2D molecular descriptors were computed from the SMILES notation of each ligand, capturing essential structural and physicochemical features. These descriptors, along with the predicted binding affinities, were used to train a deep learning model to predict binding affinity as a regression task. The model was evaluated using metrics such as Mean Squared Error (MSE), Mean Absolute Error (MAE), and R-squared. Results indicated strong performance with an R squared value of 0.861 for the training set and 0.655 for the test set, suggesting good model generalization. The model shows promise for predicting ligand-receptor interactions and can be applied in drug discovery efforts targeting PPAR-related diseases., Comment: Contain graphical abstract, 1 figure, and 3 tables

We build upon time-series classification by leveraging the capabilities of Vision Language Models (VLMs). We find that VLMs produce competitive results after two or less epochs of fine-tuning. We develop a novel approach that incorporates graphical data representations as images in conjunction with numerical data. This approach is rooted in the hypothesis that graphical representations can provide additional contextual information that numerical data alone may not capture. Additionally, providing a graphical representation can circumvent issues such as limited context length faced by LLMs. To further advance this work, we implemented a scalable end-to-end pipeline for training on different scenarios, allowing us to isolate the most effective strategies for transferring learning capabilities from LLMs to Time Series Classification (TSC) tasks. Our approach works with univariate and multivariate time-series data. In addition, we conduct extensive and practical experiments to show how this approach works for time-series classification and generative labels.

The detection of hate speech has become increasingly important in combating online hostility and its real-world consequences. Despite recent advancements, there is limited research addressing hate speech detection in Devanagari-scripted languages, where resources and tools are scarce. While large language models (LLMs) have shown promise in language-related tasks, traditional fine-tuning approaches are often infeasible given the size of the models. In this paper, we propose a Parameter Efficient Fine tuning (PEFT) based solution for hate speech detection and target identification. We evaluate multiple LLMs on the Devanagari dataset provided by (Thapa et al., 2025), which contains annotated instances in 2 languages - Hindi and Nepali. The results demonstrate the efficacy of our approach in handling Devanagari-scripted content.

Retrieving semantically relevant documents in niche domains poses significant challenges for traditional TF-IDF-based systems, often resulting in low similarity scores and suboptimal retrieval performance. This paper addresses these challenges by introducing an iterative and semi-automated query refinement methodology tailored to Humber College's career services webpages. Initially, generic queries related to interview preparation yield low top-document similarities (approximately 0.2--0.3). To enhance retrieval effectiveness, we implement a two-fold approach: first, domain-aware query refinement by incorporating specialized terms such as resources-online-learning, student-online-services, and career-advising; second, the integration of structured educational descriptors like "online resume and interview improvement tools." Additionally, we automate the extraction of domain-specific keywords from top-ranked documents to suggest relevant terms for query expansion. Through experiments conducted on five baseline queries, our semi-automated iterative refinement process elevates the average top similarity score from approximately 0.18 to 0.42, marking a substantial improvement in retrieval performance. The implementation details, including reproducible code and experimental setups, are made available in our GitHub repositories \url{https://github.com/Elipei88/HumberChatbotBackend} and \url{https://github.com/Nisarg851/HumberChatbot}. We also discuss the limitations of our approach and propose future directions, including the integration of advanced neural retrieval models., Comment: To be submitted to CoLM 2025

Retrieval-Augmented Generation (RAG) systems have become pivotal in leveraging vast corpora to generate informed and contextually relevant responses, notably reducing hallucinations in Large Language Models. Despite significant advancements, these systems struggle to efficiently process and retrieve information from large datasets while maintaining a comprehensive understanding of the context. This paper introduces SKETCH, a novel methodology that enhances the RAG retrieval process by integrating semantic text retrieval with knowledge graphs, thereby merging structured and unstructured data for a more holistic comprehension. SKETCH, demonstrates substantial improvements in retrieval performance and maintains superior context integrity compared to traditional methods. Evaluated across four diverse datasets: QuALITY, QASPER, NarrativeQA, and Italian Cuisine-SKETCH consistently outperforms baseline approaches on key RAGAS metrics such as answer_relevancy, faithfulness, context_precision and context_recall. Notably, on the Italian Cuisine dataset, SKETCH achieved an answer relevancy of 0.94 and a context precision of 0.99, representing the highest performance across all evaluated metrics. These results highlight SKETCH's capability in delivering more accurate and contextually relevant responses, setting new benchmarks for future retrieval systems., Comment: 16 pages, 8 figures, Workshop on Generative AI and Knowledge Graphs (GenAIK) at The 31st International Conference on Computational Linguistics (COLING 2025)

Large Language Model (LLM) based coding tools have been tremendously successful as software development assistants, yet they are often designed for general purpose programming tasks and perform poorly for more specialized domains such as high performance computing. Creating specialized models and tools for these domains is crucial towards gaining the benefits of LLMs in areas such as HPC. While previous work has explored HPC-specific models, LLMs still struggle to generate parallel code and it is not at all clear what hurdles are still holding back these LLMs and what must be done to overcome them. In this work, we conduct an in-depth study along the many axes of fine-tuning a specialized HPC LLM in order to better understand the challenges. Based on our findings we fine-tune and evaluate a specialized HPC LLM that is shown to be the best performing open-source code LLM for parallel code generation to date.

Gravitational Wave (GW) sources are standard sirens that provide an independent way to map the cosmic expansion history by combining with an independent redshift measurement either from an electromagnetic counterpart for a bright siren or using different statistical techniques for dark sirens. In this analysis, we perform the first Blinded Mock Data Challenge (Blinded-MDC) to test the robustness in inferring the value of Hubble constant $H_0$ for a dark siren technique which depends on astrophysical mass distribution of Binary Black Holes (BBHs). We have considered different analysis setups for the Blinded-MDC to test both statistical and systematic uncertainties and demonstrate the capabilities in inferring $H_0$ with detector sensitivity as per the fourth observation run of LIGO-Virgo-KAGRA. We find that when the astrophysical population of BBHs matches with the underlying assumption of the model, a cosmological pipeline can recover the injected parameters using the observed mass distribution. However, when the mock mass distribution of the astrophysical population depends slightly on redshift and one is ignorant about it in analyzing the data, it can cause a systematic discrepancy in the inferred value of $H_0$ by about $1.5\sigma$, above the statistical fluctuations due to noise and a limited number of events. The discrepancy in $H_0$ is arising due to astrophysical mis-modeling, and in the future, elaborate studies will be required to mitigate systematic uncertainties due to unknown astrophysical populations of BBHs. This MDC framework sets the road map for inspecting the precision and accuracy of standard siren cosmology and provides the first insight into the robustness of the population-dependent cosmology inference in a blinded analysis setup., Comment: 19 pages (+ 9 pages appendices) and 20 figures (+ 9 figures in appendices). Prepared for submission to The Astrophysical Journal

Cypher, the query language for Neo4j graph databases, plays a critical role in enabling graph-based analytics and data exploration. While substantial research has been dedicated to natural language to SQL query generation (Text2SQL), the analogous problem for graph databases referred to as Text2Cypher remains underexplored. In this work, we introduce SynthCypher, a fully synthetic and automated data generation pipeline designed to address this gap. SynthCypher employs a novel LLMSupervised Generation-Verification framework, ensuring syntactically and semantically correct Cypher queries across diverse domains and query complexities. Using this pipeline, we create SynthCypher Dataset, a large-scale benchmark containing 29.8k Text2Cypher instances. Fine-tuning open-source large language models (LLMs), including LLaMa-3.1- 8B, Mistral-7B, and QWEN-7B, on SynthCypher yields significant performance improvements of up to 40% on the Text2Cypher test set and 30% on the SPIDER benchmark adapted for graph databases. This work demonstrates that high-quality synthetic data can effectively advance the state-of-the-art in Text2Cypher tasks.

This article presents a comparative analysis of GPU-parallelized implementations of the quantum-inspired evolutionary optimization (QIEO) approach and one of the well-known classical metaheuristic techniques, the genetic algorithm (GA). The study assesses the performance of both algorithms on highly non-linear, non-convex, and non-separable function optimization problems, viz., Ackley, Rosenbrock, and Rastrigin, that are representative of the complex real-world optimization problems. The performance of these algorithms is checked by varying the population sizes by keeping all other parameters constant and comparing the fitness value it reached along with the number of function evaluations they required for convergence. The results demonstrate that QIEO performs better for these functions than GA, by achieving the target fitness with fewer function evaluations and significantly reducing the total optimization time approximately three times for the Ackley function and four times for the Rosenbrock and Rastrigin functions. Furthermore, QIEO exhibits greater consistency across trials, with a steady convergence rate that leads to a more uniform number of function evaluations, highlighting its reliability in solving challenging optimization problems. The findings indicate that QIEO is a promising alternative to GA for these kind of functions.

We study user history modeling via Transformer encoders in deep learning recommendation models (DLRM). Such architectures can significantly improve recommendation quality, but usually incur high latency cost necessitating infrastructure upgrades or very small Transformer models. An important part of user history modeling is early fusion of the candidate item and various methods have been studied. We revisit early fusion and compare concatenation of the candidate to each history item against appending it to the end of the list as a separate item. Using the latter method, allows us to reformulate the recently proposed amortized history inference algorithm M-FALCON \cite{zhai2024actions} for the case of DLRM models. We show via experimental results that appending with cross-attention performs on par with concatenation and that amortization significantly reduces inference costs. We conclude with results from deploying this model on the LinkedIn Feed and Ads surfaces, where amortization reduces latency by 30\% compared to non-amortized inference., Comment: 5 pages, 3 figures, WWW 2025

Electroencephalography (EEG) is widely researched for neural decoding in Brain Computer Interfaces (BCIs) as it is non-invasive, portable, and economical. However, EEG signals suffer from inter- and intra-subject variability, leading to poor performance. Recent technological advancements have led to deep learning (DL) models that have achieved high performance in various fields. However, such large models are compute- and resource-intensive and are a bottleneck for real-time neural decoding. Data distribution shift can be handled with the help of domain adaptation techniques of transfer learning (fine-tuning) and adversarial training that requires model parameter updates according to the target domain. One such recent technique is Parameter-efficient fine-tuning (PEFT), which requires only a small fraction of the total trainable parameters compared to fine-tuning the whole model. Therefore, we explored PEFT methods for adapting EEG-based mental imagery tasks. We considered two mental imagery tasks: speech imagery and motor imagery, as both of these tasks are instrumental in post-stroke neuro-rehabilitation. We proposed a novel ensemble of weight-decomposed low-rank adaptation methods, EDoRA, for parameter-efficient mental imagery task adaptation through EEG signal classification. The performance of the proposed PEFT method is validated on two publicly available datasets, one speech imagery, and the other motor imagery dataset. In extensive experiments and analysis, the proposed method has performed better than full fine-tune and state-of-the-art PEFT methods for mental imagery EEG classification.

In this work, we compare the domain-specific translation performance of open-source autoregressive decoder-only large language models (LLMs) with task-oriented machine translation (MT) models. Our experiments focus on the medical domain and cover four language pairs with varied resource availability: English-to-French, English-to-Portuguese, English-to-Swahili, and Swahili-to-English. Despite recent advancements, LLMs exhibit a clear gap in specialized translation quality compared to multilingual encoder-decoder MT models such as NLLB-200. In three out of four language directions in our study, NLLB-200 3.3B outperforms all LLMs in the size range of 8B parameters in medical translation. While fine-tuning LLMs such as Mistral and Llama improves their performance at medical translation, these models still fall short compared to fine-tuned NLLB-200 3.3B models. Our findings highlight the ongoing need for specialized MT models to achieve higher-quality domain-specific translation, especially in medium-resource and low-resource settings. As larger LLMs outperform their 8B variants, this also encourages pre-training domain-specific medium-sized LMs to improve quality and efficiency in specialized translation tasks.

Recent advances in self-supervised models for natural language, vision, and protein sequences have inspired the development of large genomic DNA language models (DNALMs). These models aim to learn generalizable representations of diverse DNA elements, potentially enabling various genomic prediction, interpretation and design tasks. Despite their potential, existing benchmarks do not adequately assess the capabilities of DNALMs on key downstream applications involving an important class of non-coding DNA elements critical for regulating gene activity. In this study, we introduce DART-Eval, a suite of representative benchmarks specifically focused on regulatory DNA to evaluate model performance across zero-shot, probed, and fine-tuned scenarios against contemporary ab initio models as baselines. Our benchmarks target biologically meaningful downstream tasks such as functional sequence feature discovery, predicting cell-type specific regulatory activity, and counterfactual prediction of the impacts of genetic variants. We find that current DNALMs exhibit inconsistent performance and do not offer compelling gains over alternative baseline models for most tasks, while requiring significantly more computational resources. We discuss potentially promising modeling, data curation, and evaluation strategies for the next generation of DNALMs. Our code is available at https://github.com/kundajelab/DART-Eval., Comment: NeurIPS Datasets and Benchmarks 2024

This paper presents pragmatic solutions for verifying complex mathematical algorithms implemented in hardware in an efficient and effective manner. Maximizing leverage of a known-answer-test strategy, based on predefined data scenarios combined with design-for-verification modes, we demonstrate how to find and isolate concept and design bugs early in the flow. The solutions presented are based on real project experience with single chip radar sensors for a variety of applications. The verification environments supporting the presented strategies are based on SystemVerilog and the Universal Verification Methodology., Comment: Published at DVCon Europe 2024

High-quality machine learning (ML)-ready datasets play a foundational role in developing new artificial intelligence (AI) models or fine-tuning existing models for scientific applications such as weather and climate analysis. Unfortunately, despite the growing development of new deep learning models for weather and climate, there is a scarcity of curated, pre-processed machine learning (ML)-ready datasets. Curating such high-quality datasets for developing new models is challenging particularly because the modality of the input data varies significantly for different downstream tasks addressing different atmospheric scales (spatial and temporal). Here we introduce WxC-Bench (Weather and Climate Bench), a multi-modal dataset designed to support the development of generalizable AI models for downstream use-cases in weather and climate research. WxC-Bench is designed as a dataset of datasets for developing ML-models for a complex weather and climate system, addressing selected downstream tasks as machine learning phenomenon. WxC-Bench encompasses several atmospheric processes from meso-$\beta$ (20 - 200 km) scale to synoptic scales (2500 km), such as aviation turbulence, hurricane intensity and track monitoring, weather analog search, gravity wave parameterization, and natural language report generation. We provide a comprehensive description of the dataset and also present a technical validation for baseline analysis. The dataset and code to prepare the ML-ready data have been made publicly available on Hugging Face -- https://huggingface.co/datasets/nasa-impact/WxC-Bench

Torque and continuous rotation are fundamental methods of actuation and manipulation in rigid robots. Soft robot arms use soft materials and structures to mimic the passive compliance of biological arms that bend and extend. This use of compliance prevents soft arms from continuously transmitting and exerting torques to interact with their environment. Here, we show how relying on patterning structures instead of inherent material properties allows soft robotic arms to remain compliant while continuously transmitting torque to their environment. We demonstrate a soft robotic arm made from a pair of mechanical metamaterials that act as compliant constant-velocity joints. The joints are up to 52 times stiffer in torsion than bending and can bend up to 45{\deg}. This robot arm can continuously transmit torque while deforming in all other directions. The arm's mechanical design achieves high motion repeatability (0.4 mm and 0.1{\deg}) when tracking trajectories. We then trained a neural network to learn the inverse kinematics, enabling us to program the arm to complete tasks that are challenging for existing soft robots such as installing light bulbs, fastening bolts, and turning valves. The arm's passive compliance makes it safe around humans and provides a source of mechanical intelligence, enabling it to adapt to misalignment when manipulating objects. This work will bridge the gap between hard and soft robotics with applications in human assistance, warehouse automation, and extreme environments.

This study employs an event study methodology to investigate the market impact of the U.S. Securities and Exchange Commission's (SEC) classification of crypto assets as securities. It explores how SEC interventions influence asset returns and trading volumes, focusing on explicitly named crypto assets. The empirical analysis highlights significant adverse market reactions, notably returns plummeting 12% over one week post-announcement, persisting for a month. We demonstrate that the severity of market reaction depends on sentiment and asset characteristics such as market size, age, volatility, and illiquidity. Further, we identify significant ex-ante trading volume effects indicative of pre-announcement informed trading., Comment: 20 pages, 4 tables, 2 appendix tables, 2 appendix figures

The neutron capture cross-section for the $^{181}$Ta(n,$\gamma$)$^{182}$Ta reaction has been experimentally measured at the neutron energies 0.53 and 1.05 MeV using off-line $\gamma$-ray spectrometry. $^{115}$In(n,n'$\gamma$)$^{115m}$In is used as a reference monitor reaction cross-section. The neutron was produced via the $^{7}$Li(p,n)$^{7}$Be reaction. The present study measures the cross-sections with their uncertainties and correlation matrix. The self-attenuation process, $\gamma$-ray correction factor, and low background neutron energy contribution have been calculated. The measured neutron spectrum averaged cross-sections of $^{181}$Ta(n,$\gamma$)$^{182}$Ta are discussed and compared with the existing data from the EXFOR database and also with the ENDF/B-VIII.0, TENDL-2019, JENDL-5, JEFF-3.3 evaluated data libraries.

Cancer clinical trials often face challenges in recruitment and engagement due to a lack of participant-facing informational and educational resources. This study investigated the potential of Large Language Models (LLMs), specifically GPT4, in generating patient-friendly educational content from clinical trial informed consent forms. Using data from ClinicalTrials.gov, we employed zero-shot learning for creating trial summaries and one-shot learning for developing multiple-choice questions, evaluating their effectiveness through patient surveys and crowdsourced annotation. Results showed that GPT4-generated summaries were both readable and comprehensive, and may improve patients' understanding and interest in clinical trials. The multiple-choice questions demonstrated high accuracy and agreement with crowdsourced annotators. For both resource types, hallucinations were identified that require ongoing human oversight. The findings demonstrate the potential of LLMs "out-of-the-box" to support the generation of clinical trial education materials with minimal trial-specific engineering, but implementation with a human-in-the-loop is still needed to avoid misinformation risks.

Making analogies is fundamental to cognition. Proportional analogies, which consist of four terms, are often used to assess linguistic and cognitive abilities. For instance, completing analogies like "Oxygen is to Gas as is to " requires identifying the semantic relationship (e.g., "type of") between the first pair of terms ("Oxygen" and "Gas") and finding a second pair that shares the same relationship (e.g., "Aluminum" and "Metal"). In this work, we introduce a 15K Multiple-Choice Question Answering (MCQA) dataset for proportional analogy completion and evaluate the performance of contemporary Large Language Models (LLMs) in various knowledge-enhanced prompt settings. Specifically, we augment prompts with three types of knowledge: exemplar, structured, and targeted. Our results show that despite extensive training data, solving proportional analogies remains challenging for current LLMs, with the best model achieving an accuracy of 55%. Notably, we find that providing targeted knowledge can better assist models in completing proportional analogies compared to providing exemplars or collections of structured knowledge. Our code and data are available at: https://github.com/Thiliniiw/KnowledgePrompts/, Comment: Accepted at COLING 2025

A speaker verification (SV) system offers an authentication service designed to confirm whether a given speech sample originates from a specific speaker. This technology has paved the way for various personalized applications that cater to individual preferences. A noteworthy challenge faced by SV systems is their ability to perform consistently across a range of emotional spectra. Most existing models exhibit high error rates when dealing with emotional utterances compared to neutral ones. Consequently, this phenomenon often leads to missing out on speech of interest. This issue primarily stems from the limited availability of labeled emotional speech data, impeding the development of robust speaker representations that encompass diverse emotional states. To address this concern, we propose a novel approach employing the CycleGAN framework to serve as a data augmentation method. This technique synthesizes emotional speech segments for each specific speaker while preserving the unique vocal identity. Our experimental findings underscore the effectiveness of incorporating synthetic emotional data into the training process. The models trained using this augmented dataset consistently outperform the baseline models on the task of verifying speakers in emotional speech scenarios, reducing equal error rate by as much as 3.64% relative.

Ensuring the safety of autonomous vehicles (AVs) requires both accurate estimation of their performance and efficient discovery of potential failure cases. This paper introduces Bayesian adaptive multifidelity sampling (BAMS), which leverages the power of adaptive Bayesian sampling to achieve efficient discovery while simultaneously estimating the rate of adverse events. BAMS prioritizes exploration of regions with potentially low performance, leading to the identification of novel and critical scenarios that traditional methods might miss. Using real-world AV data we demonstrate that BAMS discovers 10 times as many issues as Monte Carlo (MC) and importance sampling (IS) baselines, while at the same time generating rate estimates with variances 15 and 6 times narrower than MC and IS baselines respectively., Comment: Published at CoRL 2024: https://openreview.net/forum?id=bftFwjSJxk

Existing Large Multimodal Models (LMMs) generally focus on only a few regions and languages. As LMMs continue to improve, it is increasingly important to ensure they understand cultural contexts, respect local sensitivities, and support low-resource languages, all while effectively integrating corresponding visual cues. In pursuit of culturally diverse global multimodal models, our proposed All Languages Matter Benchmark (ALM-bench) represents the largest and most comprehensive effort to date for evaluating LMMs across 100 languages. ALM-bench challenges existing models by testing their ability to understand and reason about culturally diverse images paired with text in various languages, including many low-resource languages traditionally underrepresented in LMM research. The benchmark offers a robust and nuanced evaluation framework featuring various question formats, including true/false, multiple choice, and open-ended questions, which are further divided into short and long-answer categories. ALM-bench design ensures a comprehensive assessment of a model's ability to handle varied levels of difficulty in visual and linguistic reasoning. To capture the rich tapestry of global cultures, ALM-bench carefully curates content from 13 distinct cultural aspects, ranging from traditions and rituals to famous personalities and celebrations. Through this, ALM-bench not only provides a rigorous testing ground for state-of-the-art open and closed-source LMMs but also highlights the importance of cultural and linguistic inclusivity, encouraging the development of models that can serve diverse global populations effectively. Our benchmark is publicly available., Comment: A Multilingual Multimodal cultural benchmark for 100 languages

Edge inference techniques partition and distribute Deep Neural Network (DNN) inference tasks among multiple edge nodes for low latency inference, without considering the core-level heterogeneity of edge nodes. Further, default DNN inference frameworks also do not fully utilize the resources of heterogeneous edge nodes, resulting in higher inference latency. In this work, we propose a hierarchical DNN partitioning strategy (HiDP) for distributed inference on heterogeneous edge nodes. Our strategy hierarchically partitions DNN workloads at both global and local levels by considering the core-level heterogeneity of edge nodes. We evaluated our proposed HiDP strategy against relevant distributed inference techniques over widely used DNN models on commercial edge devices. On average our strategy achieved 38% lower latency, 46% lower energy, and 56% higher throughput in comparison with other relevant approaches., Comment: 7 pages, 8 figures, 1 table, and 1 algorithm. The manuscript is accepted to be published in 28th Design, Automation and Test in Europe Conference (IEEE DATE, 2025)

Low internal and external quantum efficiencies in high Aluminium content AlGaN-based deep-ultraviolet light-emitting diode (DUV-LED) occurred due to strong polarization effects, spontaneous and piezoelectric polarization, at the interface between two materials. It also leads to a low carrier confinement and Quantum Confined Stark Effect (QCSE), contributing to the efficiency droop of the DUV-LED. This work demonstrates an under-level MQW configuration implemented in a DUV-LED with a 257 nm emission wavelength. Three DUV-LED structures, above-, same- and under-level MQW were investigated, covering important optoelectronics properties such as energy band diagram, carrier concentrations, radiative recombination rates and electric field distribution. It is found that the quantum efficiencies, luminescence intensity and light output power of the under-level configuration has been enhanced by nine-, ten- and five-folds, relative to the above-level MQW configuration., Comment: 22 Pages, 6 Figures, 1 Table

The proliferation of AI techniques for image generation, coupled with their increasing accessibility, has raised significant concerns about the potential misuse of these images to spread misinformation. Recent AI-generated image detection (AGID) methods include CNNDetection, NPR, DM Image Detection, Fake Image Detection, DIRE, LASTED, GAN Image Detection, AIDE, SSP, DRCT, RINE, OCC-CLIP, De-Fake, and Deep Fake Detection. However, we argue that the current state-of-the-art AGID techniques are inadequate for effectively detecting contemporary AI-generated images and advocate for a comprehensive reevaluation of these methods. We introduce the Visual Counter Turing Test (VCT^2), a benchmark comprising ~130K images generated by contemporary text-to-image models (Stable Diffusion 2.1, Stable Diffusion XL, Stable Diffusion 3, DALL-E 3, and Midjourney 6). VCT^2 includes two sets of prompts sourced from tweets by the New York Times Twitter account and captions from the MS COCO dataset. We also evaluate the performance of the aforementioned AGID techniques on the VCT$^2$ benchmark, highlighting their ineffectiveness in detecting AI-generated images. As image-generative AI models continue to evolve, the need for a quantifiable framework to evaluate these models becomes increasingly critical. To meet this need, we propose the Visual AI Index (V_AI), which assesses generated images from various visual perspectives, including texture complexity and object coherence, setting a new standard for evaluating image-generative AI models. To foster research in this domain, we make our https://huggingface.co/datasets/anonymous1233/COCO_AI and https://huggingface.co/datasets/anonymous1233/twitter_AI datasets publicly available., Comment: 13 pages, 9 figures

Course Outcome (CO) and Program Outcome (PO)/Program-Specific Outcome (PSO) alignment is a crucial task for ensuring curriculum coherence and assessing educational effectiveness. The construction of a Course Articulation Matrix (CAM), which quantifies the relationship between COs and POs/PSOs, typically involves assigning numerical values (0, 1, 2, 3) to represent the degree of alignment. In this study, We experiment with four models from the BERT family: BERT Base, DistilBERT, ALBERT, and RoBERTa, and use multiclass classification to assess the alignment between CO and PO/PSO pairs. We first evaluate traditional machine learning classifiers, such as Decision Tree, Random Forest, and XGBoost, and then apply transfer learning to evaluate the performance of the pretrained BERT models. To enhance model interpretability, we apply Explainable AI technique, specifically Local Interpretable Model-agnostic Explanations (LIME), to provide transparency into the decision-making process. Our system achieves accuracy, precision, recall, and F1-score values of 98.66%, 98.67%, 98.66%, and 98.66%, respectively. This work demonstrates the potential of utilizing transfer learning with BERT-based models for the automated generation of CAMs, offering high performance and interpretability in educational outcome assessment., Comment: 26 pages, 9 figures

With the expanding use of the short-form video format in advertising, social media, entertainment, education and more, there is a need for such media to both captivate and be remembered. Video memorability indicates to us how likely a video is to be remembered by a viewer who has no emotional or personal connection with its content. This paper presents the results of using generative outpainting to expand the screen size of a short-form video with a view to improving its memorability. Advances in machine learning and deep learning are compared and leveraged to understand how extending the borders of video screensizes can affect their memorability to viewers. Using quantitative evaluation we determine the best-performing model for outpainting and the impact of outpainting based on image saliency on video memorability scores

Reliably predicting nuclear properties across the entire chart of isotopes is important for applications ranging from nuclear astrophysics to superheavy science to nuclear technology. To this day, however, all the theoretical models that can scale at the level of the chart of isotopes remain semi phenomenological. Because they are fitted locally, their predictive power can vary significantly; different versions of the same theory provide different predictions. Bayesian model mixing takes advantage of such imperfect models to build a local mixture of a set of models to make improved predictions. Earlier attempts to use Bayesian model mixing for mass table calculations relied on models treated at single-reference energy density functional level, which fail to capture some of the correlations caused by configuration mixing or the restoration of broken symmetries. In this study we have applied Bayesian model mixing techniques within a multi-reference energy density functional (MR-EDF) framework. We considered predictions of two-particle separation energies from particle number projection or angular momentum projection with four different energy density functionals - a total of eight different MR-EDF models. We used a hierarchical Bayesian stacking framework with a Dirichlet prior distribution over weights together with an inverse log-ratio transform to enable positive correlations between different models. We found that Bayesian model mixing provide significantly improved predictions over results from single MR-EDF calculations.

Existing claim verification datasets often do not require systems to perform complex reasoning or effectively interpret multimodal evidence. To address this, we introduce a new task: multi-hop multimodal claim verification. This task challenges models to reason over multiple pieces of evidence from diverse sources, including text, images, and tables, and determine whether the combined multimodal evidence supports or refutes a given claim. To study this task, we construct MMCV, a large-scale dataset comprising 15k multi-hop claims paired with multimodal evidence, generated and refined using large language models, with additional input from human feedback. We show that MMCV is challenging even for the latest state-of-the-art multimodal large language models, especially as the number of reasoning hops increases. Additionally, we establish a human performance benchmark on a subset of MMCV. We hope this dataset and its evaluation task will encourage future research in multimodal multi-hop claim verification., Comment: COLING 2025

This paper presents auto-tuned primal-dual successive convexification (Auto-SCvx), an algorithm designed to reliably achieve dynamically-feasible trajectory solutions for constrained hypersonic reentry optimal control problems across a large mission parameter space. In Auto-SCvx, we solve a sequence of convex subproblems until convergence to a solution of the original nonconvex problem. This method iteratively optimizes dual variables in closed-form in order to update the penalty hyperparameters used in the primal variable updates. A benefit of this method is that it is auto-tuning, and requires no hand-tuning by the user with respect to the constraint penalty weights. Several example hypersonic reentry problems are posed and solved using this method, and comparative studies are conducted against current methods. In these numerical studies, our algorithm demonstrates equal and often improved performance while not requiring hand-tuning of penalty hyperparameters., Comment: 38 pages, 27 figures; submitted to the AIAA Journal of Guidance, Control, and Dynamics (JGCD)

Computing photo-z for AGN is challenging, primarily due to the interplay of relative emissions associated with the SMBH and its host galaxy. SED fitting methods, effective in pencil-beam surveys, face limitations in all-sky surveys with fewer bands available, lacking the ability to capture the AGN contribution to the SED accurately. This limitation affects the many 10s of millions of AGN clearly singled out and identified by SRG/eROSITA. Our goal is to significantly enhance photometric redshift performance for AGN in all-sky surveys while avoiding the need to merge multiple data sets. Instead, we employ readily available data products from the 10th Data Release of the Imaging Legacy Survey for DESI, covering > 20,000 deg$^{2}$ with deep images and catalog-based photometry in the grizW1-W4 bands. We introduce PICZL, a machine-learning algorithm leveraging an ensemble of CNNs. Utilizing a cross-channel approach, the algorithm integrates distinct SED features from images with those obtained from catalog-level data. Full probability distributions are achieved via the integration of Gaussian mixture models. On a validation sample of 8098 AGN, PICZL achieves a variance $\sigma_{\textrm{NMAD}}$ of 4.5% with an outlier fraction $\eta$ of 5.6%, outperforming previous attempts to compute accurate photo-z for AGN using ML. We highlight that the model's performance depends on many variables, predominantly the depth of the data. A thorough evaluation of these dependencies is presented in the paper. Our streamlined methodology maintains consistent performance across the entire survey area when accounting for differing data quality. The same approach can be adopted for future deep photometric surveys such as LSST and Euclid, showcasing its potential for wide-scale realisation. With this paper, we release updated photo-z (including errors) for the XMM-SERVS W-CDF-S, ELAIS-S1 and LSS fields., Comment: Accepted for publication in Astronomy & Astrophysics. 24 pages, 21 figures

We present results from the first application of the Global Navigation Satellite System (GNSS; GPS is one example of a collection of satellites in GNSS) for radio beam calibration using a commercial GNSS receiver with the Deep Dish Development Array (D3A) at the Dominion Radio Astrophysical Observatory (DRAO). Several GNSS satellites pass through the main and side lobes of the beam each day, enabling efficient mapping of the 2D beam structure. Due to the high SNR and abundance of GNSS satellites, we find evidence that GNSS can probe several side lobes of the beam through repeatable measurements of the beam over several days. Over three days of measurements, we find a measured difference reaching a minimum of 0.56 db-Hz in the main lobe of the primary beam. These results show promise for the use of GNSS in beam mapping for the Canadian Hydrogen Observatory and Radio-transient Detector (CHORD) and other future "large-N" radio interferometers. They also motivate future development of the technique within radio astronomy., Comment: Submitted to PASA. Comments welcome!

Per-example gradient norms are a vital ingredient for estimating gradient noise scale (GNS) with minimal variance. Observing the tensor contractions required to compute them, we propose a method with minimal FLOPs in 3D or greater tensor regimes by simultaneously computing the norms while computing the parameter gradients. Using this method we are able to observe the GNS of different layers at higher accuracy than previously possible. We find that the total GNS of contemporary transformer models is predicted well by the GNS of only the normalization layers. As a result, focusing only on the normalization layer, we develop a custom kernel to compute the per-example gradient norms while performing the LayerNorm backward pass with zero throughput overhead. Tracking GNS on only those layers, we are able to guide a practical batch size schedule that reduces training time by 18% on a Chinchilla-optimal language model., Comment: 23 pages, 16 figures, to be published in the proceedings of the 2024 Conference on Neural Information Processing Systems (NeurIPS), code is available at: https://github.com/CerebrasResearch/nanoGNS

Task-oriented dialogue systems are essential for applications ranging from customer service to personal assistants and are widely used across various industries. However, developing effective multi-domain systems remains a significant challenge due to the complexity of handling diverse user intents, entity types, and domain-specific knowledge across several domains. In this work, we propose DARD (Domain Assigned Response Delegation), a multi-agent conversational system capable of successfully handling multi-domain dialogs. DARD leverages domain-specific agents, orchestrated by a central dialog manager agent. Our extensive experiments compare and utilize various agent modeling approaches, combining the strengths of smaller fine-tuned models (Flan-T5-large & Mistral-7B) with their larger counterparts, Large Language Models (LLMs) (Claude Sonnet 3.0). We provide insights into the strengths and limitations of each approach, highlighting the benefits of our multi-agent framework in terms of flexibility and composability. We evaluate DARD using the well-established MultiWOZ benchmark, achieving state-of-the-art performance by improving the dialogue inform rate by 6.6% and the success rate by 4.1% over the best-performing existing approaches. Additionally, we discuss various annotator discrepancies and issues within the MultiWOZ dataset and its evaluation system.