Workspace

Sourcing Reports

Search

Latest arXiv sourcing reports uploaded by Archibald.

Preview • weekly-arxiv-sourcing_2026-03-16_08-00.md

📚 Weekly arXiv Sourcing Digest (V1)

Generated: 2026-03-16T08:00:37.219290Z

🎯 Scope

cs.RO, cs.AI, cs.LG, cs.CV, cs.SY, cs.ET, cs.AR, cs.DC, eess.SP, eess.SY, cond-mat.mtrl-sci, physics.app-ph

📊 Coverage

MetricValue
Fetched495
Deduplicated430
Recent (<=21d)430
Scored430
High priority27
Watchlist99
Low-signal304

High Priority Leads

  1. SNAP-V: A RISC-V SoC with Configurable Neuromorphic Acceleration for Small-Scale Spiking Neural Networks

    • Score: 100 | Category: cs.AR | Company signal: no
    • Why relevant: Spiking Neural Networks (SNNs) have gained significant attention in edge computing due to their low power consumption and computational efficiency.
    • Paper: Open on arXiv

  2. Accelerating Diffusion Models for Generative AI Applications with Silicon Photonics

    • Score: 93 | Category: cs.AR | Company signal: no
    • Why relevant: Diffusion models have revolutionized generative AI, with their inherent capacity to generate highly realistic state-of-the-art synthetic data.
    • Paper: Open on arXiv

  3. FLUX: Accelerating Cross-Embodiment Generative Navigation Policies via Rectified Flow and Static-to-Dynamic Learning

    • Score: 91 | Category: cs.RO | Company signal: no
    • Why relevant: Autonomous navigation requires a broad spectrum of skills, from static goal-reaching to dynamic social traversal, yet evaluation remains fragmented across disparate protocols.
    • Paper: Open on arXiv

  4. HyperCroc: End-to-End Open-Source RISC-V MCU with a Plug-In Interface for Domain-Specific Accelerators

    • Score: 89 | Category: cs.AR | Company signal: no
    • Why relevant: Domain-Specific architectures with accelerators for machine learning and signal processing require efficient bulk data movement and high-bandwidth access to large datasets.
    • Paper: Open on arXiv

  5. TacVLA: Contact-Aware Tactile Fusion for Robust Vision-Language-Action Manipulation

    • Score: 88 | Category: cs.RO | Company signal: yes
    • Why relevant: Vision-Language-Action (VLA) models have demonstrated significant advantages in robotic manipulation.
    • Paper: Open on arXiv

  6. One-Step Flow Policy: Self-Distillation for Fast Visuomotor Policies

    • Score: 85 | Category: cs.RO | Company signal: no
    • Why relevant: Generative flow and diffusion models provide the continuous, multimodal action distributions needed for high-precision robotic policies.
    • Paper: Open on arXiv

  7. AnchorVLA4D: an Anchor-Based Spatial-Temporal Vision-Language-Action Model for Robotic Manipulation

    • Score: 81 | Category: cs.RO | Company signal: no
    • Why relevant: Since current Vision-Language-Action (VLA) systems suffer from limited spatial perception and the absence of memory throughout manipulation, we investigate visual anchors as a means to enhance spatial and temporal reason...
    • Paper: Open on arXiv

  8. COHORT: Hybrid RL for Collaborative Large DNN Inference on Multi-Robot Systems Under Real-Time Constraints

    • Score: 79 | Category: cs.DC | Company signal: no
    • Why relevant: Large deep neural networks (DNNs), especially transformer-based and multimodal architectures, are computationally demanding and challenging to deploy on resource-constrained edge platforms like field robots.
    • Paper: Open on arXiv

Watchlist

  1. Distributed Kalman--Consensus Filtering with Adaptive Uncertainty Weighting for Multi-Object Tracking in Mobile Robot Networks

    • Score: 59 | Category: eess.SY | Company signal: no
    • Why relevant: This paper presents an implementation and evaluation of a Distributed Kalman--Consensus Filter (DKCF) for Multi-Object Tracking (MOT) in mobile robot networks operating under partial observability and heterogeneous local...
    • Paper: Open on arXiv

  2. PhysMoDPO: Physically-Plausible Humanoid Motion with Preference Optimization

    • Score: 58 | Category: cs.CV | Company signal: no
    • Why relevant: Recent progress in text-conditioned human motion generation has been largely driven by diffusion models trained on large-scale human motion data.
    • Paper: Open on arXiv

  3. SortScrews: A Dataset and Baseline for Real-time Screw Classification

    • Score: 57 | Category: cs.CV | Company signal: yes
    • Why relevant: Automatic identification of screw types is important for industrial automation, robotics, and inventory management.
    • Paper: Open on arXiv

  4. Physical mechanisms of ohmic contact and tunnel diode: A novel explanation in terms of impurity-photovoltaic-effect resulting from infrared self-emission at room-temperature

    • Score: 56 | Category: physics.app-ph | Company signal: no
    • Why relevant: A mechanism of quantum-mechanical tunneling is based on electron-wavefunction and is used to explain ohmic contact and tunnel diode.
    • Paper: Open on arXiv

  5. Autocorrelation effects in a stochastic-process model for decision making via time series

    • Score: 55 | Category: cs.ET | Company signal: no
    • Why relevant: Decision makers exploiting photonic chaotic dynamics obtained by semiconductor lasers provide an ultrafast approach to solving multi-armed bandit problems by using a temporal optical signal as the driving source for sequ...
    • Paper: Open on arXiv

  6. System-Technology Co-Optimization of Bitline Routing and Bonding Pathways in Monolithic 3D DRAM Architectures

    • Score: 55 | Category: cs.AR | Company signal: no
    • Why relevant: 3D DRAM has emerged as a promising approach for continued density scaling, but its viability is limited by routing and hybrid bonding constraints to periphery, which may degrade sensing margin, latency, and array efficie...
    • Paper: Open on arXiv

  7. Beyond Imitation: Reinforcement Learning Fine-Tuning for Adaptive Diffusion Navigation Policies

    • Score: 54 | Category: cs.RO | Company signal: no
    • Why relevant: Diffusion-based robot navigation policies trained on large-scale imitation learning datasets, can generate multi-modal trajectories directly from the robot's visual observations, bypassing the traditional localization-ma...
    • Paper: Open on arXiv

  8. Learning Geometric and Photometric Features from Panoramic LiDAR Scans for Outdoor Place Categorization

    • Score: 54 | Category: cs.RO | Company signal: no
    • Why relevant: Semantic place categorization, which is one of the essential tasks for autonomous robots and vehicles, allows them to have capabilities of self-decision and navigation in unfamiliar environments.
    • Paper: Open on arXiv

  9. LLM Constitutional Multi-Agent Governance

    • Score: 54 | Category: cs.AI | Company signal: no
    • Why relevant: Large Language Models (LLMs) can generate persuasive influence strategies that shift cooperative behavior in multi-agent populations, but a critical question remains: does the resulting cooperation reflect genuine prosoc...
    • Paper: Open on arXiv

  10. InterEdit: Navigating Text-Guided Multi-Human 3D Motion Editing

  • Score: 53 | Category: cs.CV | Company signal: yes
  • Why relevant: Text-guided 3D motion editing has seen success in single-person scenarios, but its extension to multi-person settings is less explored due to limited paired data and the complexity of inter-person interactions.
  • Paper: Open on arXiv

Low-signal Sample

  1. HaltNav: Reactive Visual Halting over Lightweight Topological Priors for Robust Vision-Language Navigation

    • Score: 37 | Category: cs.RO | Company signal: no
    • Why relevant: Vision-and-Language Navigation (VLN) is shifting from rigid, step-by-step instruction following toward open-vocabulary, goal-oriented autonomy.
    • Paper: Open on arXiv

  2. Geometry-Guided Camera Motion Understanding in VideoLLMs

    • Score: 37 | Category: cs.CV | Company signal: yes
    • Why relevant: Camera motion is a fundamental geometric signal that shapes visual perception and cinematic style, yet current video-capable vision-language models (VideoLLMs) rarely represent it explicitly and often fail on fine-graine...
    • Paper: Open on arXiv

  3. Decentralized Orchestration Architecture for Fluid Computing: A Secure Distributed AI Use Case

    • Score: 37 | Category: cs.DC | Company signal: no
    • Why relevant: Distributed AI and IoT applications increasingly execute across heterogeneous resources spanning end devices, edge/fog infrastructure, and cloud platforms, often under different administrative domains.
    • Paper: Open on arXiv

  4. Unified theory of orientation averaging in X-ray spectroscopies: understanding polarization dependence in a Cartesian tensor approach

    • Score: 37 | Category: cond-mat.mtrl-sci | Company signal: no
    • Why relevant: X-ray absorption spectroscopy (XAS) and resonant inelastic X-ray scattering (RIXS) are powerful probes of electronic structure owing to their chemical and orbital selectivity.
    • Paper: Open on arXiv

  5. Scale-Plan: Scalable Language-Enabled Task Planning for Heterogeneous Multi-Robot Teams

    • Score: 36 | Category: cs.ET | Company signal: no
    • Why relevant: Long-horizon task planning for heterogeneous multi-robot systems is essential for deploying collaborative teams in real-world environments; yet, it remains challenging due to the large volume of perceptual information, m...
    • Paper: Open on arXiv

  6. Subtime: Reversible Information Exchange and the Emergence of Classical Time

    • Score: 36 | Category: cs.DC | Company signal: no
    • Why relevant: We formalize the concept of subtime -- a reversible mode of information interchange within entangled systems -- and show how classical time emerges as an asymptotic limit through decoherence.
    • Paper: Open on arXiv

  7. Maximum-Entropy Random Walks on Hypergraphs

    • Score: 36 | Category: eess.SY | Company signal: no
    • Why relevant: Random walks are fundamental tools for analyzing complex networked systems, including social networks, biological systems, and communication infrastructures.
    • Paper: Open on arXiv

  8. SldprtNet: A Large-Scale Multimodal Dataset for CAD Generation in Language-Driven 3D Design

    • Score: 35 | Category: cs.CV | Company signal: no
    • Why relevant: We introduce SldprtNet, a large-scale dataset comprising over 242,000 industrial parts, designed for semantic-driven CAD modeling, geometric deep learning, and the training and fine-tuning of multimodal models for 3D des...
    • Paper: Open on arXiv