arXiv subjects

Benjamin Meyer

Publications and source records attributed to Benjamin Meyer.

5 recordsLinked to original sources

arXiv2025

How is remifentanil dosed without dedicated indicator?

This study investigates the paradigm of intraoperative analgesic dosage using a data-driven approach based on retrospective clinical data. Remifentanil, an analgesic widely used during anesthesia, presents a dosing challenge due to the absence of an universally accepted indicator of analgesia. To examine how changes in patient state correlate with adjustments in remifentanil target concentration triggered by the practitioner, we analyzed data from two sources: VitalDB (Seoul, Korea) and PREDIMED (Grenoble, France). Results show that only features derived from arterial pressure are consistently associated with changes in remifentanil targets. This finding is robust across both datasets despite variations in specific thresholds. In particular, increases in remifentanil targets are associated with high or rising arterial pressure over short periods (1--2 minutes), whereas decreases are linked to low, stable, or declining arterial pressure over longer periods (5--7 minutes). By capturing anesthesiologists' dosing strategies we provide a foundation for the future development of closed-loop control algorithms. Beyond the specific example of remifentanil's change prediction, the proposed feature generation and associated sparse fitting approach can be applied to other domain where human decision can be viewed as sensors interpretation.

eess.SP↗

arXiv2025

A document is worth a structured record: Principled inductive bias design for document recognition

Many document types use intrinsic, convention-driven structures that serve to encode precise and structured information, such as the conventions governing engineering drawings. However, many state-of-the-art approaches treat document recognition as a mere computer vision problem, neglecting these underlying document-type-specific structural properties, making them dependent on sub-optimal heuristic post-processing and rendering many less frequent or more complicated document types inaccessible to modern document recognition. We suggest a novel perspective that frames document recognition as a transcription task from a document to a record. This implies a natural grouping of documents based on the intrinsic structure inherent in their transcription, where related document types can be treated (and learned) similarly. We propose a method to design structure-specific relational inductive biases for the underlying machine-learned end-to-end document recognition systems, and a respective base transformer architecture that we successfully adapt to different structures. We demonstrate the effectiveness of the so-found inductive biases in extensive experiments with progressively complex record structures from monophonic sheet music, shape drawings, and simplified engineering drawings. By integrating an inductive bias for unrestricted graph structures, we train the first-ever successful end-to-end model to transcribe mechanical engineering drawings to their inherently interlinked information. Our approach is relevant to inform the design of document recognition systems for document types that are less well understood than standard OCR, OMR, etc., and serves as a guide to unify the design of future document foundation models.

cs.CV↗

arXiv2025

A Comprehensive Survey of Agents for Computer Use: Foundations, Challenges, and Future Directions

Agents for computer use (ACUs) are an emerging class of systems capable of executing complex tasks on digital devices -- such as desktops, mobile phones, and web platforms -- given instructions in natural language. These agents can automate tasks by controlling software via low-level actions like mouse clicks and touchscreen gestures. However, despite rapid progress, ACUs are not yet mature for everyday use. In this survey, we investigate the state-of-the-art, trends, and research gaps in the development of practical ACUs. We provide a comprehensive review of the ACU landscape, introducing a unifying taxonomy spanning three dimensions: (I) the domain perspective, characterizing agent operating contexts; (II) the interaction perspective, describing observation modalities (e.g., screenshots, HTML) and action modalities (e.g., mouse, keyboard, code execution); and (III) the agent perspective, detailing how agents perceive, reason, and learn. We review 87 ACUs and 33 datasets across foundation model-based and classical approaches through this taxonomy. Our analysis identifies six major research gaps: insufficient generalization, inefficient learning, limited planning, low task complexity in benchmarks, non-standardized evaluation, and a disconnect between research and practical conditions. To address these gaps, we advocate for: (a) vision-based observations and low-level control to enhance generalization; (b) adaptive learning beyond static prompting; (c) effective planning and reasoning methods and models; (d) benchmarks that reflect real-world task complexity; (e) standardized evaluation based on task success; (f) aligning agent design with real-world deployment constraints. Together, our taxonomy and analysis establish a foundation for advancing ACU research toward general-purpose agents for robust and scalable computer use.

cs.AI↗

arXiv2020

Methods for Actors in the Electric Power System to Prevent, Detect and React to ICT Attacks and Failures

The fundamental changes in power supply and increasing decentralization require more active grid operation and an increased integration of ICT at all power system actors. This trend raises complexity and increasingly leads to interactions between primary grid operation and ICT as well as different power system actors. For example, virtual power plants control various assets in the distribution grid via ICT to jointly market existing flexibilities. Failures of ICT or targeted attacks can thus have serious effects on security of supply and system stability. This paper presents a holistic approach to providing methods specifically for actors in the power system for prevention, detection, and reaction to ICT attacks and failures. The focus of our measures are solutions for ICT monitoring, systems for the detection of ICT attacks and intrusions in the process network, and the provision of actionable guidelines as well as a practice environment for the response to potential ICT security incidents.

cs.CR↗

arXiv2020

Learning the energy curvature versus particle number in approximate density functionals

The average energy curvature as a function of the particle number is a molecule-specific quantity, which measures the deviation of a given functional from the exact conditions of density functional theory (DFT). Related to the lack of derivative discontinuity in approximate exchange-correlation potentials, the information about the curvature has been successfully used to restore the physical meaning of Kohn-Sham orbital eigenvalues and to develop non-empirical tuning and correction schemes for density functional approximations. In this work, we propose the construction of a machine-learning framework targeting the average energy curvature between the neutral and the radical cation state of thousands of small organic molecules (QM7 database). The applicability of the model is demonstrated in the context of system-specific gamma-tuning of the LC-$ω$PBE functional and validated against the molecular first ionization potentials at equation-of-motion (EOM) coupled-cluster references. In addition, we propose a local version of the non-linear regression model and demonstrate its transferability and predictive power by determining the optimal range-separation parameter for two large molecules relevant to the field of hole-transporting materials. Finally, we explore the underlying structure of the QM7 database with the t-SNE dimensionality-reduction algorithm and identify structural and compositional patterns that promote the deviation from the piecewise linearity condition.

physics.chem-ph↗