BioSecBench-Surveillance: A Verifiable Benchmark for AI Agents in Pathogen Genomic Surveillance

Authors: Harmon Bhasin, Kevin Flyangolts, Dianzhuo Wang, Evan Seeyave, Arjun Banerjee, Amanda Darling, Joshua Stallings, David Stern, Shawn Higdon, Claire Duvallet, Bryan Tegomoh, Kenny Workman

Abstract: As pathogen genomic surveillance scales, the bottleneck is shifting from data generation to analysis. We present BioSecBench-Surveillance, a verifiable benchmark of 100 evaluations testing whether AI agents can infer the right analysis pipeline from raw sequencing data and surveillance context. Each evaluation gives an agent only the data and context a human analyst would have, then grades its str… ▽ More As pathogen genomic surveillance scales, the bottleneck is shifting from data generation to analysis. We present BioSecBench-Surveillance, a verifiable benchmark of 100 evaluations testing whether AI agents can infer the right analysis pipeline from raw sequencing data and surveillance context. Each evaluation gives an agent only the data and context a human analyst would have, then grades its structured answer deterministically. The tasks span seven categories, from taxonomic classification to genetic-engineering detection, across diverse sample types and sequencing technologies. Across 3,962 gradable attempts from sixteen model-harness pairs, the strongest configuration cleared only about half. Opus 4.8 with PI led at 50.2 percent, with a 95 percent confidence interval of 40.1 to 60.3 percent across 83 evaluations, tied with GPT-5.5 with Codex at 50.2 percent, with a 95 percent confidence interval of 40.8 to 59.6 percent, followed by Opus 4.7 with PI at 49.6 percent, with a 95 percent confidence interval of 40.0 to 59.2 percent, and Sonnet 4.6 with PI at 48.6 percent, with a 95 percent confidence interval of 38.9 to 58.3 percent. Even when agents invoked the correct workflows, their mistakes came from the choices around them, such as which references, thresholds, filters, and normalization to apply. BioSecBench-Surveillance provides a standard for measuring whether agents can be trusted to perform genomic surveillance when the next outbreak arrives. △ Less

Submitted 21 July, 2026; originally announced July 2026.

A framework to model real-time databases

Authors: Nizar Idoudi, Nada louati, Claude Duvallet, Bruno Sadeg, Rafik Bouaziz, Faiez Gargouri

Abstract: Real-time databases deal with time-constrained data and time-constrained transactions. The design of this kind of databases requires the introduction of new concepts to support both data structures and the dynamic behaviour of the database. In this paper, we give an overview about different aspects of real-time databases and we clarify requirements of their modelling. Then, we present a framewor… ▽ More Real-time databases deal with time-constrained data and time-constrained transactions. The design of this kind of databases requires the introduction of new concepts to support both data structures and the dynamic behaviour of the database. In this paper, we give an overview about different aspects of real-time databases and we clarify requirements of their modelling. Then, we present a framework for real-time database design and describe its fundamental operations. A case study demonstrates the validity of the structural model and illustrates SQL queries and Java code generated from the classes of the model △ Less

Submitted 8 January, 2010; originally announced January 2010.

A New Approach to Manage QoS in Distributed Multimedia Systems

Authors: Bechir Alaya, Claude Duvallet, Bruno Sadeg

Abstract: Dealing with network congestion is a criterion used to enhance quality of service (QoS) in distributed multimedia systems. The existing solutions for the problem of network congestion ignore scalability considerations because they maintain a separate classification for each video stream. In this paper, we propose a new method allowing to control QoS provided to clients according to the network c… ▽ More Dealing with network congestion is a criterion used to enhance quality of service (QoS) in distributed multimedia systems. The existing solutions for the problem of network congestion ignore scalability considerations because they maintain a separate classification for each video stream. In this paper, we propose a new method allowing to control QoS provided to clients according to the network congestion, by discarding some frames when needed. The technique proposed, called (m,k)-frame, is scalable with little degradation in application performances. (m,k)-frame method is issued from the notion of (m,k)-firm realtime constraints which means that among k invocations of a task, m invocations must meet their deadline. Our simulation studies show the usefulness of (m,k)-frame method to adapt the QoS to the real conditions in a multimedia application, according to the current system load. Notably, the system must adjust the QoS provided to active clients1 when their number varies, i.e. dynamic arrival of clients. △ Less

Submitted 26 June, 2009; originally announced June 2009.