benchmark dataset
PulseAugur coverage of benchmark dataset — every cluster mentioning benchmark dataset across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
PoseOFF improves human action anticipation for robots · 2 sources tracked
Researchers have developed PoseOFF, a novel pose-anchored optical flow representation designed to improve human action anticipation in human-robot interaction. This method captures localized motion around human joints, …
-
New geometric framework analyzes robustness of AI fairness audits
Researchers have developed a new geometric framework to analyze the robustness of neighborhood-based fairness audits in machine learning. These audits assess individual fairness by comparing predictions for similar indi…
-
HNDiff framework integrates atmospheric physics for advanced image dehazing · 2 sources tracked
Researchers have developed Haze-Noise Diffusion (HNDiff), a novel diffusion framework for image dehazing that incorporates the atmospheric scattering model. This approach grounds diffusion in physical principles, ensuri…
-
EgoGVAE reconstructs full-body meshes from head pose data
Researchers have developed EgoGVAE, a novel method for reconstructing full-body meshes from head pose data, which is crucial for applications using head-mounted devices and smartglasses. Unlike previous diffusion-based …
-
Zero-Fi uses signal-language alignment for zero-shot Wi-Fi activity recognition
Researchers have developed Zero-Fi, a novel framework for Wi-Fi-based human activity recognition that utilizes contrastive signal-language alignment. This approach allows the system to recognize new activities without n…
-
New decoding method preserves LLM refusal behavior under high-temperature sampling
Researchers have developed a new decoding method called Refusal-Gated Decoding to maintain an LLM's safety guardrails when using high-temperature sampling. This technique aims to increase output diversity without sacrif…
-
Object perception enhances single-view 3D reconstruction in new research
Researchers have developed a novel method to improve single-view 3D object reconstruction by integrating object perception signals. This approach leverages pretrained perception models to extract semantic and geometric …
-
New HopS method improves prompt learning for vision-language models with partial labels
Researchers have developed a new method called Holistic Optimal Label Selection (HopS) to improve prompt learning for vision-language models when only partial labels are available. HopS employs two strategies: a local f…
-
Physics-guided LLM framework achieves 98.49% accuracy in bearing fault diagnosis
Researchers have developed a novel physics-guided framework that leverages large language models for bearing fault diagnosis. This system addresses challenges in feature efficiency, traceability to fault physics, and mu…
-
New dataset integrates electric potential for improved ECT image reconstruction
Researchers have developed a new benchmark dataset for Electrical Capacitance Tomography (ECT) image reconstruction that incorporates electric potential fields. This dataset, generated using a COMSOL-MATLAB pipeline, in…
-
Paper calls for LLM benchmarks resistant to pretraining data contamination
A new paper argues that benchmark datasets used to evaluate large language models (LLMs) must be resistant to contamination from pretraining data. The authors highlight that many current benchmarks are already included …