Existing fine-grained evaluation approaches for retrieval-augmented generation (RAG) such as RAGAS, ARES, or RagChecker can provide insights into the performance of various steps of the RAG pipeline, but could not detect failures beyond a predefined set of errors. At the same time, a simple approach of manually inspecting a small subset of predictions is helpful to discover arbitrary error types, but requires human effort and lacks scalability. In this work, we test the possibility of automating per-example inspection of predictions with the use of modern large language models. We prompt GPT-4o to analyse RAG failure cases in an open-ended fashion and then ask it to output a structured summarized report of discovered error types. In our case study, we find that automatically generated error reports can capture main error types listed in the human expert’s reports and that automatic per-example analysis is correct in 80% of cases. At the same time, we notice occasional issues with correctly interpreting the particular failure case, over-generalization of error types in the report summary step and some variability in reports when relaunching the analysis.
Open-ended error analysis in retrieval-augmented generation
| Nadezhda Chirkova |
| Workshop on Evaluation of RAG Systems (Eval4RAG), co-located at the 47th European Conference on Information Retrieval (ECIR), Lucca, Italy, 10 April, 2025 |
To make robots autonomous in real-world everyday spaces, they should be able to learn from their interactions within these spaces, how to best execute tasks specified by non-expert users in a safe and reliable way. To do so requires sequential decision-making skills that combine machine learning, adaptive planning and control in uncertain environments as well as solving hard combinatorial optimisation problems. Our research combines expertise in reinforcement learning, computer vision, robotic control, sim2real transfer, large multimodal foundation models and neural combinatorial optimisation to build AI-based architectures and algorithms to improve robot autonomy and robustness when completing everyday complex tasks in constantly changing environments.
The research we conduct on expressive visual representations is applicable to visual search, object detection, image classification and the automatic extraction of 3D human poses and shapes that can be used for human behavior understanding and prediction, human-robot interaction or even avatar animation. We also extract 3D information from images that can be used for intelligent robot navigation, augmented reality and the 3D reconstruction of objects, buildings or even entire cities.
Our work covers the spectrum from unsupervised to supervised approaches, and from very deep architectures to very compact ones. We’re excited about the promise of big data to bring big performance gains to our algorithms but also passionate about the challenge of working in data-scarce and low-power scenarios.
Furthermore, we believe that a modern computer vision system needs to be able to continuously adapt itself to its environment and to improve itself via lifelong learning. Our driving goal is to use our research to deliver embodied intelligence to our users in robotics, autonomous driving, via phone cameras and any other visual means to reach people wherever they may be.
This web site uses cookies for the site search, to display videos and for aggregate site analytics.
Learn more about these cookies in our privacy notice.
Cookie settings
You may choose which kind of cookies you allow when visiting this website. Click on "Save cookie settings" to apply your choice.
This content is currently blocked. To view the content please either 'Accept social media cookies' or 'Accept all cookies'.
For more information on cookies see our privacy notice.