arXiv subjects

Rebecca Steinert

Publications and source records attributed to Rebecca Steinert.

9 recordsLinked to original sources

arXiv2026

SEGRA: A Structured Experience Guided Reasoning Agent for Property Graph Question Answering

Enterprise IT support knowledge graphs capture rich relationships among cases, users, devices, symptoms, taxonomic categories, root causes, and historical resolutions. Yet querying them in Gremlin requires knowledge of graph schemas, traversal semantics, edge directionality, and property-graph-specific constraints, making them difficult for non-expert operators to use. We introduce SEGRA, an experience-guided agent for enterprise text-to-Gremlin question answering. SEGRA integrates intent routing, schema- and taxonomy-grounded query generation, multi-shot decomposition, execution-aware verification, and a curriculum-bootstrapped skill library that reuses verified query patterns. On an enterprise IT support benchmark, SEGRA achieves a $7.0\times$ higher mean judge score than backbone-only chain-of-thought prompting. Its skill library further reduces LLM calls by $20\%$ and dollar cost by $18\%$ relative to SEGRA without skills, while preserving answer quality. These results show that schema-grounded agent design and reusable execution experience improve both accuracy and efficiency for enterprise graph QA.

cs.AI↗

arXiv2026

DQA: Diagnostic Question Answering for IT Support

Enterprise IT support interactions are fundamentally diagnostic: effective resolution requires iterative evidence gathering from ambiguous user reports to identify an underlying root cause. While retrieval-augmented generation (RAG) provides grounding through historical cases, standard multi-turn RAG systems lack explicit diagnostic state and therefore struggle to accumulate evidence and resolve competing hypotheses across turns. We introduce DQA, a diagnostic question-answering framework that maintains persistent diagnostic state and aggregates retrieved cases at the level of root causes rather than individual documents. DQA combines conversational query rewriting, retrieval aggregation, and state-conditioned response generation to support systematic troubleshooting under enterprise latency and context constraints. We evaluate DQA on 150 anonymized enterprise IT support scenarios using a replay-based protocol. Averaged over three independent runs, DQA achieves a 78.7% success rate under a trajectory-level success criterion, compared to 41.3% for a multi-turn RAG baseline, while reducing average turns from 8.4 to 3.9.

cs.CL↗

arXiv2026

VIGIL: Towards Edge-Extended Agentic AI for Enterprise IT Support

Enterprise IT support is constrained by heterogeneous devices, evolving policies, and long-tail failure modes that are difficult to resolve centrally. We present VIGIL, an edge-extended agentic AI system that deploys desktop-resident agents to perform situated diagnosis, retrieval over enterprise knowledge, and policy-governed remediation directly on user devices with explicit consent and end-to-end observability. In a 10-week pilot of VIGIL's operational loop on 100 resource-constrained endpoints, VIGIL reduces interaction rounds by 39%, achieves at least 4 times faster diagnosis, and supports self-service resolution in 82% of matched cases. Users report excellent usability, high trust, and low cognitive workload across four validated instruments, with qualitative feedback highlighting transparency as critical for trust. Notably, users rated the system higher when no historical matches were available, suggesting on-device diagnosis provides value independent of knowledge base coverage. This pilot establishes safety and observability foundations for fleet-wide continuous improvement.

cs.AI↗

arXiv2020

Learning Combinatorial Optimization on Graphs: A Survey with Applications to Networking

Existing approaches to solving combinatorial optimization problems on graphs suffer from the need to engineer each problem algorithmically, with practical problems recurring in many instances. The practical side of theoretical computer science, such as computational complexity, then needs to be addressed. Relevant developments in machine learning research on graphs are surveyed for this purpose. We organize and compare the structures involved with learning to solve combinatorial optimization problems, with a special eye on the telecommunications domain and its continuous development of live and research networks.

cs.LG↗

arXiv2017

Service Provider DevOps

Although there is consensus that software defined networking and network functions virtualization overhaul service provisioning and deployment, the community still lacks a definite answer on how carrier-grade operations praxis needs to evolve. This article presents what lies beyond the first evolutionary steps in network management, identifies the challenges in service verification, observability, and troubleshooting, and explains how to address them using our Service Provider DevOps (SP-DevOps) framework. We compendiously cover the entire process from design goals to tool realization and employ an elastic version of an industry-standard use case to show how on-the-fly verification, software-defined monitoring, and automated troubleshooting of services reduce the cost of fault management actions. We assess SP-DevOps with respect to key attributes of software-defined telecommunication infrastructures both qualitatively and quantitatively, and demonstrate that SP-DevOps paves the way toward carrier-grade operations and management in the network virtualization era.

cs.NI↗

arXiv2016

Final Service Provider DevOps concept and evaluation

This report presents the results of the UNIFY Service Provider DevOps activities. First, we present the final definition and assessment of the concept. SP-DevOps is realized by a combination of various functional components facilitating integrated service verification, efficient and programmable observability, and automated troubleshooting processes. Our assessment shows that SP-DevOps can help providers to reach a medium level of DevOps maturity and allows significant reduction in OPEX. Second, we focus on the evaluation of the proposed SP-DevOps components. The set of tools proposed supports ops and devs across all stages, with a focus on the deployment, operation and debugging phases, and allows to activate automated processes for operating NFV environments. Finally, we present use-cases and our demonstrators for selected process implementions, which allowed the functional validation of SP-DevOps.

cs.NI↗

arXiv2015

Initial Service Provider DevOps concept, capabilities and proposed tools

This report presents a first sketch of the Service Provider DevOps concept including four major management processes to support the roles of both service and VNF developers as well as the operator in a more agile manner. The sketch is based on lessons learned from a study of management and operational practices in the industry and recent related work with respect to management of SDN and cloud. Finally, the report identifies requirements for realizing SP-DevOps within an combined cloud and transport network environment as outlined by the UNIFY NFV architecture.

cs.NI↗

arXiv2015

Service Provider DevOps network capabilities and tools

This report provides an understanding of how the UNIFY Service Provider (SP)-DevOps concept can be applied and integrated with a combined cloud and transport network NFV architecture. Specifically, the report contains technical descriptions of a set of novel SP-DevOps tools and support functions that facilitate observability, troubleshooting, verification, and VNF development processes. The tools and support functions are described in detail together with their architectural mapping, giving a wider understanding of the SP-DevOps concept as a whole, and how SP-DevOps tools can be used for supporting orchestration and programmability in the UNIFY NFV framework. The concept is further exemplified in a case study for deployment and scaling of an Elastic Firewall.

cs.NI↗

arXiv2013

Research Directions in Network Service Chaining

Network Service Chaining (NSC) is a service deployment concept that promises increased flexibility and cost efficiency for future carrier networks. NSC has received considerable attention in the standardization and research communities lately. However, NSC is largely undefined in the peer-reviewed literature. In fact, a literature review reveals that the role of NSC enabling technologies is up for discussion, and so are the key research challenges lying ahead. This paper addresses these topics by motivating our research interest towards advanced dynamic NSC and detailing the main aspects to be considered in the context of carrier-grade telecommunication networks. We present design considerations and system requirements alongside use cases that illustrate the advantages of adopting NSC. We detail prominent research challenges during the typical lifecycle of a network service chain in an operational telecommunications network, including service chain description, programming, deployment, and debugging, and summarize our security considerations. We conclude this paper with an outlook on future work in this area.

cs.NI↗