arXiv subjects

Lev Sorokin

Publications and source records attributed to Lev Sorokin.

13 recordsLinked to original sources

arXiv2026

Search-based Testing of Vision Language Models for In-Car Scene Understanding

In the automotive domain, in-car scene understanding (ISU) enables the detection of safety-critical events, such as driver distraction, and supports drivers or passengers by analyzing the in-car scene and adapting the environment (e.g., ambient lighting). The industry is increasingly exploring vision-language models (VLMs) to interpret camera-recorded in-car scenes and extract information for downstream reasoning tasks. However, VLMs may generate incomplete, erroneous, or misleading scene descriptions, highlighting the need for systematic testing. Collecting real in-vehicle data is costly, difficult to scale, and often infeasible, particularly in early design stages. In this paper, we present ISU-Test, an automated testing approach that combines rendering-based scene generation with search-based testing to evaluate ISU systems. By framing testing as an optimization problem and systematically modifying scene parameters, our method generates diverse in-car scenarios and explores a wide range of configurations. We evaluate ISU-Test on both an industrial prototype and open-source VLMs across two case studies: question answering and captioning, comparing against randomized scenario generation. Results show that ISU-Test significantly outperforms the baseline, achieving up to 10 times higher failure rates and up to 3.6 times higher failure coverage.

cs.CV↗

arXiv2026

DeepTest Tool Competition 2026: Benchmarking an LLM-Based Automotive Assistant

This report summarizes the results of the first edition of the Large Language Model (LLM) Testing competition, held as part of the DeepTest workshop at ICSE 2026. Four tools competed in benchmarking an LLM-based car manual information retrieval application, with the objective of identifying user inputs for which the system fails to appropriately mention warnings contained in the manual. The testing solutions were evaluated based on their effectiveness in exposing failures and the diversity of the discovered failure-revealing tests. We report on the experimental methodology, the competitors, and the results.

cs.AI↗

arXiv2026

STELLAR: A Search-Based Testing Framework for Large Language Model Applications

Large Language Model (LLM)-based applications are increasingly deployed across various domains, including customer service, education, and mobility. However, these systems are prone to inaccurate, fictitious, or harmful responses, and their vast, high-dimensional input space makes systematic testing particularly challenging. To address this, we present STELLAR, an automated search-based testing framework for LLM-based applications that systematically uncovers text inputs leading to inappropriate system responses. Our framework models test generation as an optimization problem and discretizes the input space into stylistic, content-related, and perturbation features. Unlike prior work that focuses on prompt optimization or coverage heuristics, our work employs evolutionary optimization to dynamically explore feature combinations that are more likely to expose failures. We evaluate STELLAR on three LLM-based conversational question-answering systems. The first focuses on safety, benchmarking both public and proprietary LLMs against malicious or unsafe prompts. The second and third target navigation, using an open-source and an industrial retrieval-augmented system for in-vehicle venue recommendations. Overall, STELLAR exposes up to 4.3 times (average 2.5 times) more failures than the existing baseline approaches.

cs.SE↗

arXiv2025

Benchmarking Contextual Understanding for In-Car Conversational Systems

In-Car Conversational Question Answering (ConvQA) systems significantly enhance user experience by enabling seamless voice interactions. However, assessing their accuracy and reliability remains a challenge. This paper explores the use of Large Language Models (LLMs) alongside advanced prompting techniques and agent-based methods to evaluate the extent to which ConvQA system responses adhere to user utterances. The focus lies on contextual understanding and the ability to provide accurate venue recommendations considering user constraints and situational context. To evaluate utterance-response coherence using an LLM, we synthetically generate user utterances accompanied by correct and modified failure-containing system responses. We use input-output, chain-of-thought, self-consistency prompting, and multi-agent prompting techniques with 13 reasoning and non-reasoning LLMs of varying sizes and providers, including OpenAI, DeepSeek, Mistral AI, and Meta. We evaluate our approach on a case study involving restaurant recommendations. The most substantial improvements occur for small non-reasoning models when applying advanced prompting techniques, particularly multi-agent prompting. However, reasoning models consistently outperform non-reasoning models, with the best performance achieved using single-agent prompting with self-consistency. Notably, DeepSeek-R1 reaches an F1-score of 0.99 at a cost of 0.002 USD per request. Overall, the best balance between effectiveness and cost-time efficiency is reached with the non-reasoning model DeepSeek-V3. Our findings show that LLM-based evaluation offers a scalable and accurate alternative to traditional human evaluation for benchmarking contextual understanding in ConvQA systems.

cs.CL↗

arXiv2025

Automated Factual Benchmarking for In-Car Conversational Systems using Large Language Models

In-car conversational systems bring the promise to improve the in-vehicle user experience. Modern conversational systems are based on Large Language Models (LLMs), which makes them prone to errors such as hallucinations, i.e., inaccurate, fictitious, and therefore factually incorrect information. In this paper, we present an LLM-based methodology for the automatic factual benchmarking of in-car conversational systems. We instantiate our methodology with five LLM-based methods, leveraging ensembling techniques and diverse personae to enhance agreement and minimize hallucinations. We use our methodology to evaluate CarExpert, an in-car retrieval-augmented conversational question answering system, with respect to the factual correctness to a vehicle's manual. We produced a novel dataset specifically created for the in-car domain, and tested our methodology against an expert evaluation. Our results show that the combination of GPT-4 with the Input Output Prompting achieves over 90 per cent factual correctness agreement rate with expert evaluations, other than being the most efficient approach yielding an average response time of 4.5s. Our findings suggest that LLM-based testing constitutes a viable approach for the validation of conversational systems regarding their factual correctness.

cs.CL↗

arXiv2025

Simulator Ensembles for Trustworthy Autonomous Driving Systems Testing

Scenario-based testing with driving simulators is extensively used to identify failing conditions of automated driving assistance systems (ADAS). However, existing studies have shown that repeated test execution in the same as well as in distinct simulators can yield different outcomes, which can be attributed to sources of flakiness or different implementations of the physics. In this paper, we present MultiSim, a novel approach to multi-simulation ADAS testing based on a search-based testing approach that leverages an ensemble of simulators to identify failure-inducing, simulator-agnostic test scenarios. During the search, each scenario is evaluated jointly on multiple simulators. Scenarios that produce consistent results across simulators are prioritized for further exploration, while those that fail on only a subset of simulators are given less priority, as they may reflect simulator-specific issues rather than generalizable failures. Our empirical study, which involves testing three lane-keeping ADAS on different pairs of three widely used simulators, demonstrates that MultiSim outperforms single-simulator testing by achieving, on average, a higher rate of simulator-agnostic failures by 66%. Compared to a state-of-the-art multi-simulator approach that combines the outcome of independent test generation campaigns obtained in different simulators, MultiSim identifies, on average, up to 3.4X more simulator-agnostic failing tests and higher failure rates. To avoid the costly execution of test inputs on which simulators disagree, we propose to predict simulator disagreements and bypass test executions. Our results show that utilizing a surrogate model during the search retains the average number of valid failures and also improves efficiency. Our findings indicate that combining an ensemble of simulators is a promising approach for the automated cross-replication in ADAS testing.

cs.SE↗

arXiv2024

Can Search-Based Testing with Pareto Optimization Effectively Cover Failure-Revealing Test Inputs?

Search-based software testing (SBST) is a widely adopted technique for testing complex systems with large input spaces, such as Deep Learning-enabled (DL-enabled) systems. Many SBST techniques focus on Pareto-based optimization, where multiple objectives are optimized in parallel to reveal failures. However, it is important to ensure that identified failures are spread throughout the entire failure-inducing area of a search domain and not clustered in a sub-region. This ensures that identified failures are semantically diverse and reveal a wide range of underlying causes. In this paper, we present a theoretical argument explaining why testing based on Pareto optimization is inadequate for covering failure-inducing areas within a search domain. We support our argument with empirical results obtained by applying two widely used types of Pareto-based optimization techniques, namely NSGA-II (an evolutionary algorithm) and OMOPSO (a swarm-based Pareto-optimization algorithm), to two DL-enabled systems: an industrial Automated Valet Parking (AVP) system and a system for classifying handwritten digits. We measure the coverage of failure-revealing test inputs in the input space using a metric that we refer to as the Coverage Inverted Distance quality indicator. Our results show that NSGA-II-based search and OMOPSO are not more effective than a na\"ive random search baseline in covering test inputs that reveal failures. The replication package for this study is available in a GitHub repository.

cs.SE↗

arXiv2024

Towards Continuous Assurance Case Creation for ADS with the Evidential Tool Bus

An assurance case has become an integral component for the certification of safety-critical systems. While manually defining assurance case patterns can be not avoided, system-specific instantiations of assurance case patterns are both costly and time-consuming. It becomes especially complex to maintain an assurance case for a system when the requirements of the System-Under-Assurance change, or an assurance claim becomes invalid due to, e.g., degradation of a systems component, as common when deploying learning-enabled components. In this paper, we report on our preliminary experience leveraging the tool integration framework Evidential Tool Bus (ETB) for the construction and continuous maintenance of an assurance case from a predefined assurance case pattern. Specifically, we demonstrate the assurance process on an industrial Automated Valet Parking system from the automotive domain. We present the formalization of the provided assurance case pattern in the ETB processable logical specification language of workflows. Our findings show that ETB is able to create and maintain evidence required for the construction of an assurance case.

cs.SE↗

arXiv2024

Guiding the Search Towards Failure-Inducing Test Inputs Using Support Vector Machines

In this paper, we present NSGA-II-SVM (Non-dominated Sorting Genetic Algorithm with Support Vector Machine Guidance), a novel learnable evolutionary and search-based testing algorithm that leverages Support Vector Machine (SVM) classification models to direct the search towards failure-revealing test inputs. Supported by genetic search, NSGA-II-SVM creates iteratively SVM-based models of the test input space, learning which regions in the search space are promising to be explored. A subsequent sampling and repetition of evolutionary search iterations allow to refine and make the model more accurate in the prediction. Our preliminary evaluation of NSGA-II-SVM by testing an Automated Valet Parking system shows that NSGA-II-SVM is more effective in identifying more critical test cases than a state of the art learnable evolutionary testing technique as well as naive random search.

cs.SE↗

arXiv2023

Monitoring Auditable Claims in the Cloud

When deploying mission-critical systems in the cloud, where deviations may have severe consequences, the assurance of critical decisions becomes essential. Typical cloud systems are operated by third parties and are built on complex software stacks consisting of e.g., Kubernetes, Istio, or Kafka, which due to their size are difficult to be verified. Nevertheless, one needs to make sure that mission-critical choices are made correctly. We propose a flexible runtime monitoring approach that is independent of the implementation of the observed system that allows to monitor safety and data-related properties. Our approach is based on combining distributed Datalog-based programs with tamper-proof storage based on Trillian to verify the premises of safety-critical actions. The approach can be seen as a generalization of the Certificate Transparency project. We apply our approach to an industrial use case that uses a cloud infrastructure for orchestrating unmanned air vehicles.

cs.SE↗

arXiv2023

OpenSBT: A Modular Framework for Search-based Testing of Automated Driving Systems

Search-based software testing (SBT) is an effective and efficient approach for testing automated driving systems (ADS). However, testing pipelines for ADS testing are particularly challenging as they involve integrating complex driving simulation platforms and establishing communication protocols and APIs with the desired search algorithm. This complexity prevents a wide adoption of SBT and thorough empirical comparative experiments with different simulators and search approaches. We present OpenSBT, an open-source, modular and extensible framework to facilitate the SBT of ADS. With OpenSBT, it is possible to integrate simulators with an embedded system under test, search algorithms and fitness functions for testing. We describe the architecture and show the usage of our framework by applying different search algorithms for testing Automated Emergency Braking Systems in CARLA as well in the high-fidelity Prescan simulator in collaboration with our industrial partner DENSO. OpenSBT is available at https://git.fortiss.org/opensbt. A demo video is provided here: https://youtu.be/6csl\_UAOD\_4.

cs.SE↗

arXiv2023

Reflections on Surrogate-Assisted Search-Based Testing: A Taxonomy and Two Replication Studies based on Industrial ADAS and Simulink Models

Surrogate-assisted search-based testing (SA-SBT) aims to reduce the computational time for testing compute-intensive systems. Surrogates enhance testing techniques by improving test case generation focusing the testing budget on the most critical portions of the input domain. In addition, they can serve as approximations of the system under test (SUT) to predict tests' results instead of executing the tests on compute-intensive SUTs. This article reflects on the existing SA-SBT techniques, particularly those applied to system-level testing and often facilitated using simulators or complex test beds. Our objective is to synthesize different heuristic algorithms and evaluation methods employed in existing SA-SBT techniques and present a comprehensive view of SA-SBT solutions. In addition, by critically reviewing our previous work on SA-SBT, we aim to identify the limitations in our proposed algorithms and evaluation methods and to propose potential improvements. We present a taxonomy that categorizes and contrasts existing SA-SBT solutions and highlights key research gaps. To identify the evaluation challenges, we conduct two replication studies of our past SA-SBT solutions: One study uses industrial advanced driver assistance system (ADAS) and the other relies on a Simulink model benchmark. We compare our results with those of the original studies and identify the difficulties in evaluating SA-SBT techniques, including the impact of different contextual factors on results generalization and the validity of our evaluation metrics. Based on our taxonomy and replication studies, we propose future research directions, including re-considerations in the current evaluation metrics used for SA-SBT solutions, utilizing surrogates for fault localization and repair in addition to testing, and creating frameworks for large-scale experiments by applying SA-SBT to multiple SUTs and simulators.

cs.SE↗

arXiv2022

Towards Auditable Distributed Systems

The emerging trend towards distributed (cloud) systems (DS) has widely arrived whether in the automotive, public or the financial sector, but the execution of services of heterogeneous service providers is exposed to several risks. Beside hardware/software faults or cyber attacks that can influence the correctness of the system, fraud is also an issue. In such case it is not only important to verify the correctness of the system, but also have evidence which component and participant behaves faulty. This makes it possible, e.g. to claim for compensation after systems execution but also to assure information for verification can be trusted. The main goal of our research is to assure the monitoring of DS based on auditable information. We follow a decentralized monitoring strategy and envision a distributed monitoring approach of system properties based on distributedlogic programs that consider auditability. The expected contribution of this work is to establish with the application of our framework the mutual trust of distributed parties, as well as trust of clients in the systems execution. We showcase our ideas on a DS for booking services with unmanned air vehicles.

cs.SE↗