Modern database systems are increasingly expected to support hybrid queries that combine traditional relational operators with vector-based similarity search enabling applications such as semantic search, contextual recommendations, and multimodal analytics. However, existing benchmarks target either purely relational workloads or isolated similarity search, leaving a gap in evaluating database systems that aim to integrate both. In this paper, we present RVBench , a benchmarking framework for evaluating hybrid relational–vector workloads. RVBench comes with a relational data model that adapts the real-world MediaWiki schema to include vector columns and provides a suite of parameterized SQL templates that interleave similarity search with traditional SQL operations. The framework supports multiple similarity semantics (neighbors based on top-$k$, rank intervals, and sampled ranks) and allows generating hybrid workloads to evaluate database systems on both performance and retrieval quality. We demonstrate RVBench by benchmarking and comparing PostgreSQL and other commercial databases with a reference implementation of the workloads to highlight key trade-offs in query execution strategies and accuracy.

Comprehending long videos remains a significant challenge for Large Multi-modal Models (LMMs). Current LMMs struggle to process even minutes to hours videos due to their lack of explicit memory and retrieval mechanisms. To address this limitation, we propose RAVU (Retrieval Augmented Video Understanding), a novel framework for video understanding enhanced by retrieval with compositional reasoning over a spatio-temporal graph. We construct a graph representation of the video, capturing both spatial and temporal relationships between entities. This graph serves as a long-term memory, allowing us to track objects and their actions across time. To answer complex queries, we decompose the queries into a sequence of reasoning steps and execute these steps on the graph, retrieving relevant key information. Our approach enables more accurate understanding of long videos, particularly for queries that require multi-hop reasoning and tracking objects across frames. Our approach demonstrate superior performances with limited retrieved frames (5-10) compared with other SOTA methods and baselines on two major video QA datasets: NExT-QA and EgoSchema.

Reinforcement learning from human feedback (RLHF) has emerged as the primary method for aligning large language models (LLMs) with human preferences.While it enables LLMs to achieve human-level alignment, it often incurs significant computational and financial costs due to its reliance on training external reward models or human-labeled preferences.In this work, we propose Implicit Preference Optimization (IPO), an alternative approach that leverages generative LLMs as preference classifiers, thereby reducing the dependence on external human feedback or reward models to obtain preferences.We conduct a comprehensive evaluation on the preference classification ability of LLMs using RewardBench, assessing models across different sizes, architectures, and training levels to validate our hypothesis.Furthermore, we investigate the self-improvement capabilities of LLMs by generating multiple responses for a given instruction and employing the model itself as a preference classifier for Direct Preference Optimization (DPO)-based training.Our findings demonstrate that models trained through IPO achieve performance comparable to those utilizing state-of-the-art reward models for obtaining preferences.Our code is available at https://github.com/shivank21/ Implicit_Preference_Optimization.

Graph Neural Networks (Gnns) have achieved remarkable success in various downstream tasks, such as node classification and link prediction. Yet, efficiently deploying Gnns remains a challenge due to their computational complexity. Graph knowledge distillation aims to address this by transferring task-specific structural knowledge from teacher Gnns to lightweight student Gnns or Multi-Layer Perceptrons (MLPs). Despite its promise, existing distillation approaches suffer from several limitations: (i) they require extensive task-specific supervision(ii) they must be retrained separately for each downstream task, and (iii) they often struggle in heterophilous settings. To overcome these challenges, we propose TAG2M, a Task-Agnostic Gnn-to-MLP distillation framework designed for efficient and accurate few-shot inference. TAG2M introduces several novel strategies, including a self-supervised contrastive loss that captures topological information solely from node attributes. Additionally, it leverages Lipschitz embeddings to encode positional information with provable distortion bounds, ensuring robust representation learning. To further enhance adaptability for few-shot inference, TAG2M incorporates a learnable prompt head, which facilitates rapid task adaptation even in label-scarce settings. Unlike prior methods, TAG2M generalizes well across both homophilous and heterophilous datasets while delivering a significant computational advantage, achieving up to a 20X -200X speed-up. Extensive evaluations on 11 public datasets demonstrate its superior accuracy across diverse tasks, including node classification, link prediction, and node regression, outperforming state-of-the-art approaches.

ABSTRACT Internet of Drones (IoD) is one of the most beneficial and has many versatile applications like Surveillance and Security, Delivery and Logistics, Environmental Monitoring, Agriculture, and so forth. The IoD network is crucial for collecting sensitive data like geo‐coordinates, vehicle traffic data, and property details while surveying the various deployment locations in smart cities. The communication between users and drones can be compromised over insecure wireless channels by multiple attacks such as Man‐in‐the‐middle‐attack, Denial of Service, and so forth. Many schemes have already been propounded in the field of IoD. Still, many of them cannot address the resource constraints problem of drones, and existing protocols have higher computation and communication costs. Therefore, this paper has proposed a robust and efficient Hyper‐Elliptic Curve‐based authentication scheme (REHAS), which provides a session key for secure communication. Artificial Identities are generated using a hash function and random numbers. Fuzzy Extractor is used for user biometric authentication, which makes the smart device secure when lost. HECC is used with a smaller bit size of 80 bits rather than ECC of 160 bits. The security of the REHAS has been ensured using Scyther simulation. Furthermore, the resilience, safety, and robustness of REHAS are ensured by Informal security analysis. Lastly, a comparative study of the REHAS has been performed with other related Authentication and key agreement (AKA) protocols regarding communication cost, Computation cost, and security features, demonstrating that REHAS incurred less computation cost (6.7171 ms), communication overhead (1696 bits), and energy consumption (22.5 mJ) than other existing AKA schemes.

In this work, we focus on label efficient learning for video action detection. We develop a novel semi-supervised active learning approach which utilizes both labeled as well as un- labeled data along with informative sample selection for ac- tion detection. Video action detection requires spatio-temporal localization along with classification, which poses several challenges for both active learning (informative sample se- lection) as well as semi-supervised learning (pseudo label generation). First, we propose NoiseAug, a simple augmenta- tion strategy which effectively selects informative samples for video action detection. Next, we propose fft-attention, a novel technique based on high-pass filtering which enables effective utilization of pseudo label for SSL in video action detection by emphasizing on relevant activity region within a video. We evaluate the proposed approach on three different bench- mark datasets, UCF-101-24, JHMDB-21, and Youtube-VOS. First, we demonstrate its effectiveness on video action detec- tion where the proposed approach outperforms prior works in semi-supervised and weakly-supervised learning along with several baseline approaches in both UCF101-24 and JHMDB- 21. Next, we also show its effectiveness on Youtube-VOS for video object segmentation demonstrating its generalization capability for other dense prediction tasks in videos.

Training and evaluating language models increasingly requires the construction of meta-datasets -- diverse collections of curated data with clear provenance. Natural language prompting has recently lead to improved zero-shot generalization by transforming existing, supervised datasets into a variety of novel instruction tuning tasks, highlighting the benefits of meta-dataset curation. While successful in general-domain text, translating these data-centric approaches to biomedical language modeling remains challenging, as labeled biomedical datasets are significantly underrepresented in popular data hubs. To address this challenge, we introduce BigBio a community library of 126+ biomedical NLP datasets, currently covering 13 task categories and 10+ languages. BigBio facilitates reproducible meta-dataset curation via programmatic access to datasets and their metadata, and is compatible with current platforms for prompt engineering and end-to-end few/zero shot language model evaluation. We discuss our process for task schema harmonization, data auditing, contribution guidelines, and outline two illustrative use cases: zero-shot evaluation of biomedical prompts and large-scale, multi-task learning. BigBio is an ongoing community effort and is available at https://github.com/bigscience-workshop/biomedical

Fetal magnetic resonance imaging (MRI) is challenged by uncontrollable, large, and irregular fetal movements. It is, therefore, performed through visual monitoring of fetal motion and repeated acquisitions to ensure diagnostic-quality images are acquired. Nevertheless, visual monitoring of fetal motion based on displayed slices, and navigation at the level of stacks-of-slices is inefficient. The current process is highly operator-dependent, increases scanner usage and cost, and significantly increases the length of fetal MRI scans which makes them hard to tolerate for pregnant women. To help build automatic MRI motion tracking and navigation systems to overcome the limitations of the current process and improve fetal imaging, we have developed a new real-time image-based motion tracking method based on deep learning that learns to predict fetal motion directly from acquired images. Our method is based on a recurrent neural network, composed of spatial and temporal encoder-decoders, that infers motion parameters from anatomical features extracted from sequences of acquired slices. We compared our trained network on held-out test sets (including data with different characteristics, e.g. different fetuses scanned at different ages, and motion trajectories recorded from volunteer subjects) with networks designed for estimation as well as methods adopted to make predictions. The results show that our method outperformed alternative techniques, and achieved real-time performance with average errors of 3.5 and 8 degrees for the estimation and prediction tasks, respectively. Our real-time deep predictive motion tracking technique can be used to assess fetal movements, to guide slice acquisitions, and to build navigation systems for fetal MRI.

With the emerging advancements in VANETs, preserving location privacy has become a vital requirement. Most of the privacy schemes focus on changing the pseudonyms after a certain interval of time or in some special scenarios (can be mix-context or mix-zone) to avoid linking messages. Due to pseudonym linking attacks, elementary methods of pseudonym changing are found inefficient to provide desirable protection or privacy. The spatiotemporal information contained in beacons makes vehicles traceable, and the driver's privacy is breached. In this paper, we propose a mechanism CCAPS, which is a combination of two previously proposed schemes CAPS and CPN. This scheme takes the advantages of these two mix-context schemes and combines them. We evaluate this scheme against a global multi-target tracking adversary using PREXT simulator, and we compare the performance of proposed mechanism against the CAPS and CPN schemes.