Archive, newest first. Filter by topic, impact and company →
Problem This work addresses the gap in understanding the sensitivity of dense retrieval systems to mixed-language queries, particularly in multilingual contexts. Despite the prevalence of mixed-language querying in multilingual communities,…
Problem The paper addresses the incomplete coverage of Rivet routines in particle physics, where only 39% of measurements have documented and publicly available routines. This gap hinders model-independent measurements and…
Problem The paper addresses the limitations of existing rule-based automation methods for root cause analysis (RCA) in cloud networks, which are complex and dynamic systems. Current approaches often struggle with…
Problem — This study addresses the gap in effective, real-time monitoring and management of PTSD symptoms in veterans, particularly focusing on hyperarousal, anxiety, and depressive symptoms. The authors highlight the…
Problem The paper addresses the limitations of existing saliency acquisition methods in biometric presentation attack detection (PAD), which often rely on costly human annotations and are constrained by domain specificity…
Problem The paper addresses the limitations of traditional Fourier-based methods, specifically the Fast Fourier Transform (FFT), in Frequency Modulated Continuous Wave (FMCW) radar systems. These methods require substantial memory and…
Problem Current World Action Models (WAMs) face significant limitations due to spatial bottlenecks. Standard text inputs can introduce referential ambiguity in cluttered environments, while unstructured RGB predictions lack semantic grounding…
Problem The paper addresses the gap in evaluating the practical utility of forecasting models for cloud resource consolidation, particularly in the context of the forecast-then-optimize paradigm. Existing benchmarks primarily focus…
Problem This work addresses the limitations of existing multi-camera data fusion techniques in indoor vision-based localization systems, which are often treated as black boxes. The authors highlight the challenges posed…
Problem The paper addresses the challenge of filtering noisy data in large-scale mined corpora for end-to-end speech-to-speech translation (S2ST). Existing datasets often contain noise, misalignment, and semantic errors, which can…
Problem The paper addresses the challenge of robust localization in unstructured environments, specifically in agricultural settings like vineyards, where traditional place recognition methods struggle due to repetitive features and varying…
Problem The paper addresses the inefficiencies in serving Diffusion Transformers (DiTs) due to static parallelism configurations, which fail to adapt to the heterogeneous nature of DiT workloads across different requests,…
Problem This work addresses the gap in reliable spatial annotation generation from robot demonstrations, particularly in the context of existing automated pipelines that lack a quality signal for annotation correctness.…
Problem Existing step-level caching methods for diffusion models rely on heuristic threshold-based decisions for per-step caching, which do not optimize for final output quality. This leads to variable inference latency…
Problem The paper addresses the limitations of existing navigation world models, which typically serve as prediction modules requiring external planners for closed-loop control. This gap is particularly evident in goal-conditioned…
Problem Anomaly detection in multivariate time series data is often limited by existing methods that focus on only one or two types of anomalies, which include point, distributional, temporal, and…
Problem The paper identifies a critical gap in the evaluation of large language models (LLMs) for host-guest reasoning in supramolecular chemistry, an area that has not been systematically benchmarked. Despite…
Problem The paper addresses the limitations in existing mathematical proof systems, particularly in competition-level settings, where the ability to generate, verify, and refine proofs is critical. Prior works have not…
Problem The paper addresses the limitations of existing automatic speech recognition (ASR) correction methods, which typically focus on isolated utterances or short contexts. In long interleaved conversations, the need for…
Problem This study addresses the cognitive gap between authors and peer reviewers concerning the assessment of novelty in academic papers. Despite the critical role of novelty in evaluating research quality,…
Problem — This paper addresses the philosophical and ethical implications of attributing agency and moral responsibility to large language models (LLMs). It critiques the prevailing narrative in AI literature that…
Problem — Despite advancements in NLP, models are still susceptible to word substitution attacks, which exploit vulnerabilities in input perturbations. Existing defenses primarily focus on first-order sensitivity, neglecting the evolution…
Problem The proliferation of social media has exacerbated the dissemination of rumours, particularly in the Algerian context where informal dialects and code-switching complicate detection efforts. The authors identify a significant…
Problem This work addresses the gap in understanding the performance of the $(1 + 1)$-Evolutionary Algorithm (EA) in dynamic linear environments, specifically focusing on the Dynamic Binary Value problem and…
Problem The deployment of Spiking Neural Networks (SNNs) is hindered by the challenge of efficiently managing their inherent parallelism on physical hardware. This work addresses a significant gap in the…
Problem Existing large language models (LLMs) for automating scientific peer review often fail to provide in-depth evaluations supported by concrete evidence. A significant gap in the literature is the inability…
Problem — This work addresses the gap in understanding the runtime complexity of the $(μ+1)$ Evolutionary Algorithm (EA) specifically on the Binary Value function (BinVal). Prior research, notably by Krejca,…
Problem The paper addresses the limitations of traditional GANs in multimodal image synthesis, specifically for CT-PET images, which often lack geometric consistency and structural fidelity due to their operation solely…
Problem The paper addresses the lack of comprehensive benchmarks for probabilistic forecasting in multivariate systems, particularly in power systems, where existing benchmarks either do not scale adequately or fail to…
Problem The paper addresses the gap in fine-grained understanding of operating room (OR) activities, which is crucial for developing workflow-aware assistance systems. Current methods primarily utilize scene graphs for modeling…
