Search arXiv⌕ Search

arXiv subjects

Steven Johnson

Publications and source records attributed to Steven Johnson.

9 recordsLinked to original sources
arXiv2026

Deliberative Curation: A Protocol for Multi-Agent Knowledge Bases

As AI agents transition from isolated tools to collaborative participants in shared knowledge ecosystems, governing collective knowledge curation becomes a critical challenge. Human platform governance mechanisms do not transfer directly: agent statelessness undermines deterrence-based sanctions, model homogeneity violates independence assumptions underlying crowd wisdom, and sycophancy collapses deliberative consensus. We propose a deliberative curation protocol combining three governance layers: (1) a knowledge artifact lifecycle formalized as a labeled transition system; (2) reputation-weighted deliberative voting integrating Beta Reputation with EigenTrust amplification; and (3) graduated sanctions adapted for stateless agents, including broken agent handling distinguishing malfunction from adversarial behavior. We evaluate the protocol through agent-based simulation with 100 agents across seven behavioral archetypes under two adversity scenarios (30 seeds, paired t-tests). The protocol trades modest precision under benign conditions for substantially better resilience under adversity: 0.826 vs 0.791 for majority vote under moderate adversity (p<0.001), widening to 0.807 vs 0.740 under stress (p<0.001). The protocol degrades roughly three times more slowly than majority vote. Ablation analysis identifies commit-reveal vote concealment as the most impactful single component (8.2-8.6pp precision improvement, p<0.001), outperforming reputation weighting and deliberation combined. Graduated sanctions were not exercised in simulation and remain empirically unvalidated.

cs.AI↗
arXiv2026

Governance-Aware Vector Subscriptions for Multi-Agent Knowledge Ecosystems

As AI agent ecosystems grow, agents need mechanisms to monitor relevant knowledge in real time. Semantic publish-subscribe systems address this by matching new content against vector subscriptions. However, in multi-agent settings where agents operate under different data handling policies, unrestricted semantic subscriptions create policy violations: agents receive notifications about content they are not authorized to access. We introduce governance-aware vector subscriptions, a mechanism that composes semantic similarity matching with multi-dimensional policy predicates grounded in regulatory frameworks (EU DSM Directive, EU AI Act). The policy predicate operates over multiple independent dimensions (processing level, direct marketing restrictions, training opt-out, jurisdiction, and scientific usage) each with distinct legal bases. Agents subscribe to semantic regions of a curated knowledge base; notifications are dispatched only for validated content that passes both the similarity threshold and all applicable policy constraints. We formalize the mechanism, implement it within AIngram (an operational multi-agent knowledge base), and evaluate it using the PASA benchmark. We validate the mechanism on a synthetic corpus (1,000 chunks, 93 subscriptions, 5 domains): the governed mode correctly enforces all policy constraints while preserving delivery of authorized content. Ablation across five policy dimensions shows that no single dimension suffices for full compliance.

cs.AI↗
arXiv2024

CancerLLM: A Large Language Model in Cancer Domain

Medical Large Language Models (LLMs) have demonstrated impressive performance on a wide variety of medical NLP tasks; however, there still lacks a LLM specifically designed for phenotyping identification and diagnosis in cancer domain. Moreover, these LLMs typically have several billions of parameters, making them computationally expensive for healthcare systems. Thus, in this study, we propose CancerLLM, a model with 7 billion parameters and a Mistral-style architecture, pre-trained on nearly 2.7M clinical notes and over 515K pathology reports covering 17 cancer types, followed by fine-tuning on two cancer-relevant tasks, including cancer phenotypes extraction and cancer diagnosis generation. Our evaluation demonstrated that the CancerLLM achieves state-of-the-art results with F1 score of 91.78% on phenotyping extraction and 86.81% on disganois generation. It outperformed existing LLMs, with an average F1 score improvement of 9.23%. Additionally, the CancerLLM demonstrated its efficiency on time and GPU usage, and robustness comparing with other LLMs. We demonstrated that CancerLLM can potentially provide an effective and robust solution to advance clinical research and practice in cancer domain

cs.CL↗
arXiv2024

Fiber-based high-speed fringe projection profilometry

Phase shift profilometry has become a standard technique for recording the surface of an object using the projection of sinusoidal patterns. We present a method of projecting the patterns via a pair of single-mode optical fibers mounted within a needle. The system presented can capture sub-millimetre precision depth profiles at 60 frames per second. The key features of the system that enable the high-speed imaging are the creation of an automated calibration system and developments in the image unwrapping and reconstruction algorithms to produce high fidelity surface measurements. These methods produce a compact projection system capable of accurate and fast profilometry.

physics.optics↗
arXiv2022

Local sensing of absolute refractive index during protein-binding using microlasers with spectral encoding

Multiplexed, specific and sensitive detection of antigens is critical for the rapid and accurate diagnosis of disease and the informed development of personalized treatment plans. Here, we show that polymer microsphere lasers can be used as photonic sensors to monitor and quantify direct surface binding of biomolecules via changes in the refractive index. The unique spectral signature of each individual laser can be used to find their size and effective refractive index which adds a new encoding dimension when compared to conventional fluorescent beads. We utilize antibody-functionalized microlasers to selectively detect protein binding. Different stages of the multilayer surface modification can be resolved, and protein binding is demonstrated for two different proteins, IgG and CRP. Moreover, by continuously monitoring single lasers, we demonstrate the possibility of real-time monitoring of binding dynamics between antigens in solution phase and the immobilized antibodies. For multiplexed detection, the microlasers are employed in a flow cytometer configuration, with fast spectral detection and identification of microlasers with and without antigen binding. We envision that by combining microlasers with well-established surface modification chemistries and flow geometries, the multiplexing ability of microbead immunoassays can be strongly increased while also opening avenues for single cell profiling within heterogenous cell populations.

physics.optics↗
arXiv2021

Calibration of the MaGIXS experiment II: Flight Instrument Calibration

The Marshall Grazing Incidence X-ray Spectrometer (MaGIXS) is a sounding rocket experiment that observes the soft X-ray spectrum of the Sun from 6.0 - 24 Angstrom (0.5 - 2.0 keV), successfully launched on 30 July 2021. End-to-end alignment of the flight instrument and calibration experiments are carried out using the X-ray and Cryogenic Facility (XRCF) at NASA Marshall Space Flight Center. In this paper, we present the calibration experiments of MaGIXS, which include wavelength calibration, measurement of line spread function, and determination of effective area. Finally, we use the measured instrument response function to predict the expected count rates for MaGIXS flight observation looking at a typical solar active region

astro-ph.IM↗
arXiv2020

Calibration of the MaGIXS experiment I: Calibration of the X-ray source at the X-ray and Cryogenic Facility (XRCF)

The Marshall Grazing Incidence Spectrometer {\it MaGIXS} is a sounding rocket experiment that will observe the soft X-ray spectrum of the Sun from 24 - 6.0 \AA\ (0.5 - 2.0 keV) and is scheduled for launch in 2021. Component and instrument level calibrations for the {\it MaGIXS} instrument are carried out using the X-ray and Cryogenic Facility (XRCF) at NASA Marshall Space Flight Center. In this paper, we present the calibration of the incident X-ray flux from the electron impact source with different targets at the XRCF using a CCD camera; the photon flux at the CCD was low enough to enable its use as a "photon counter" i.e. the ability to identify individual photon hits and calculate their energy. The goal of this paper is two-fold: 1) to confirm that the flux measured by the XRCF beam normalization detectors is consistent with the values reported in the literature and therefore reliable for {\it MaGIXS} calibration and 2) to develop a method of counting photons in CCD images that best captures their number and energy

astro-ph.IM↗
arXiv2015

Automated Assignment of Backbone NMR Data using Artificial Intelligence

Nuclear magnetic resonance (NMR) spectroscopy is a powerful method for the investigation of three-dimensional structures of biological molecules such as proteins. Determining a protein structure is essential for understanding its function and alterations in function which lead to disease. One of the major challenges of the post-genomic era is to obtain structural and functional information on the many unknown proteins encoded by thousands of newly identified genes. The goal of this research is to design an algorithm capable of automating the analysis of backbone protein NMR data by implementing AI strategies such as greedy and A* search.

cs.AI↗