RETRIEVAL · RECEIPT · MEASURED 2026-10-02
arXiv technical papers
Technical papers from arXiv. Questions and scores only: the text belongs to its authors.
589 documents · 39,953 passages · 200 questions scored
BEFORE AND AFTER · SAME QUESTIONS AT EVERY STEP
WHAT MOVED THE NUMBER.
01 · NDCG@10 0.867
Cross-encoder reranker
A cross-encoder rereads the top candidates against the question and reorders them.
02 · NDCG@10 0.917
Document identity in every passage
Each passage carries a one-line identity of the paper it came from.
METHOD · AND ITS LIMIT
HOW IT IS MEASURED.
200 questions per corpus, each naming its subject, one labelled passage per question, scored on the product's retrieval receipt; nDCG@10 with a bootstrap 95% interval; questions are model-written and so easier than human test sets: use these to compare configurations, not as an absolute claim.
20 OF 200 QUESTIONS · PICKED BY HASH, NOT BY HAND
THE QUESTIONS, WITH THEIR RESULTS.
A fixed sample across the results: passages ranked first, passages found lower in the ten, and misses. This corpus shows questions and scores only. The papers are their authors' work, so no passage text is reproduced here.
In the paper on an LLM-enabled behavior-driven hardware development workflow by Müller et al., how much better does the proposed workflow perform compared to existing LLM-based approaches regarding design correctness and assertion coverage?
Labelled document: LLM-enabled behavior-driven development workflow for formally verified hardware designs — arXiv preprint — 14 Sep 2026
According to the report "Benchmarking the computational power of quantum computers," how does detecting leakage errors affect the throughput and maximum achievable capability scores on the Helios-1 device?
Labelled document: Quantum computer benchmarking — Research paper — 2026
According to the study on the 2019 Kincade Fire's impact on vineyards, what statistically significant difference in greenness recovery was observed between the areas inside and outside the burn perimeter by 2021?
Labelled document: Vineyard wildfire resilience during the 2019 Kincade Fire — Academic research paper — 14 September 2026
According to the neural network noise variability study by Alsinan et al., why does Swin-U's performance deteriorate when exposed to elevated noise levels, and how can this issue be addressed?
Labelled document: S. Alsinan et al. — arXiv preprint on neural network robustness to input noise variability — 13 Sep 2026
What constraint limits the maximum number of distinct planar wavefronts that the multi-reflector CATR measurement system described by Rohde & Schwarz in the "Emulation of Closely Spaced Spatial Sensing Targets" paper can create simultaneously?
Labelled document: Jiang et al. — arXiv preprint on ISAC OTA testing — 14 Sep 2026
What specific hardware target, development tool version, and operating frequency were used to evaluate the MXFP4 accelerator core across different hardware construction languages?
Labelled document: Daniele Passaretti et al. — Special Session Paper on HCLs and MXFP4 Accelerators — 2026
In the CoRef-GS research paper, what are the specific numerical results reported for the ReLaGS benchmark method under PnP Pose Estimation?
Labelled document: CoRef-GS multi-agent scene understanding method — arXiv preprint — 17 Sep 2026
In the "Foundation Models for Generalizable Semantic and Goal-Oriented Communication" paper, what specific parameter settings are used for the LoRA configuration applied to the Stage-B U-Net?
Labelled document: Boliang Liu et al. — arXiv preprint on FMSGOC communication framework — 7 Sep 2026
What is the primary goal of the Cnuas emulation platform, given that it prioritizes software experimentation over matching hardware execution speeds?
Labelled document: Cnuas emulation platform — arXiv preprint — 14 Sep 2026
In the extended paper on leader-follower formation control under bearing persistence of excitation, what condition ensures that the equilibrium state $( \tilde{p}_{ij}, \tilde{v}_{ij} ) = (0,0)$ achieves uniform exponential stability for a follower node $i$ and its neighbor $j$?
Labelled document: Tarek Bouazza et al. — arXiv preprint on leader-follower formation control — 16 Sep 2026
In the study on through-thickness coupling in bilayer microcantilevers by Zhou et al., what determines the thickness at which the transient overshoot disappears, and why can it shift?
Labelled document: Zhou, Minami, and Yoshikawa — arXiv preprint on microcantilever bending physics — 9 Sep 2026
In the Journal of Lightwave Technology paper on energy-efficient hollow-core fibre transmission, how do the minimum energy per bit and maximum throughput vary with span length and single-mode fibre attenuation over a 1000 km link?
Labelled document: Hollow-core fibre transmission energy efficiency — academic journal article — 2026
In the references section of the first-principles study on V2B MBene anodes by Abia and Zubair, what article by B. Liang and colleagues covers N-functionalized Ti2B MBene as a Na-ion battery anode?
Labelled document: N-functionalized V2B MBene anode study — first-principles research paper — 10 Sep 2026
In the research paper "Round-the-Clock Sub-Ambient Cooling via Dynamic Sky View Factor Steering," how is the ground radiance absorbed by the emitter calculated?
Labelled document: DSVFS cooling system — research paper — 2026
In the "From Time to Channels" paper on steel wire rope inspection, what limitation do defect-feature-oriented approaches like TFO face when trying to identify local flaws using temporal or axial signatures?
Labelled document: Siyu You et al. — arXiv preprint on steel wire rope flaw detection — 14 Sep 2026
In the research paper on the variable stiffness actuator using a crossed four-bar mechanism, by what factor is the total mass of the torque-only optimized design larger than that of the compromise design?
Labelled document: Variable stiffness actuator based on crossed four-bar mechanism — arXiv preprint — 16 Sep 2026
What maximum percentage increase in coincidence rates did researchers at the Technical University of Munich achieve in their paper on subcircuit compilation for linear optical quantum computing?
Labelled document: Tobias Forster et al. — arXiv preprint on subcircuit compilation for linear optical quantum computing — 11 Sep 2026
In the "Volumetric Evanescent Edge Coupling for Fiber-to-Chip Optical I/O" paper, how can packaging challenges caused by the system's strict angular alignment tolerance be addressed without requiring a dedicated on-chip monitor for every input channel?
Labelled document: Elshehaby et al. — research paper on volumetric evanescent edge coupling for fiber-to-chip optical I/O — 2026
Which organization provided financial support for the research documented in IEEE publication 10.1109/XXXX.2022.1234567?
Labelled document: Transparent Identity Verification Approach Using MPC and Efficient Credential Status Handling — IEEE paper — 12 Sep 2026
In the UAV-assisted 3D ISAC system study by Jiang et al., what specific UAV location represents the exact Pareto-optimal placement set when dealing with a single volume of interest?
Labelled document: Pareto-Optimal Rate-CRLB Tradeoff via Anchor Placement Optimization for UAV-Assisted 3D ISAC — Research paper — 14 Sep 2026