Adriana Correia

Maria Paz Oliva, Adriana Correia, Ivan Vankov et al.

Evaluating Natural Language Generation (NLG) is crucial for the practical adoption of AI, but has been a longstanding research challenge. While human evaluation is considered the de-facto standard, it is expensive and lacks scalability. Practical applications have driven the development of various automatic evaluation metrics (AEM), designed to compare the model output with human-written references, generating a score which approximates human judgment. Over time, AEMs have evolved from simple lexical comparisons, to semantic similarity models and, more recently, to LLM-based evaluators. However, it seems that no single metric has emerged as a definitive solution, resulting in studies using different ones without fully considering the implications. This paper aims to show this by conducting a thorough examination of the methodologies of existing metrics, their documented strengths and limitations, validation methods, and correlations with human judgment. We identify several key challenges: metrics often capture only specific aspects of text quality, their effectiveness varies by task and dataset, validation practices remain unstructured, and correlations with human judgment are inconsistent. Importantly, we find that these challenges persist in the most recent type of metric, LLM-as-a-Judge, as well as in the evaluation of Retrieval Augmented Generation (RAG), an increasingly relevant task in academia and industry. Our findings challenge the quest for the 'perfect metric'. We propose selecting metrics based on task-specific needs and leveraging complementary evaluations and advocate that new metrics should focus on enhanced validation methodologies.

Ivan Vankov, Matyo Ivanov, Adriana Correia et al.

Large Language Models (LLMs) have demonstrated considerable success in open-book question answering (QA), where the task requires generating answers grounded in a provided external context. A critical challenge in open-book QA is to ensure that model responses are based on the provided context rather than its parametric knowledge, which can be outdated, incomplete, or incorrect. Existing evaluation methods, primarily based on the LLM-as-a-judge approach, face significant limitations, including biases, scalability issues, and dependence on costly external systems. To address these challenges, we propose a novel metric that contrasts the perplexity of the model response under two conditions: when the context is provided and when it is not. The resulting score quantifies the extent to which the model's answer relies on the provided context. The validity of this metric is demonstrated through a series of experiments that show its effectiveness in identifying whether a given answer is grounded in the provided context. Unlike existing approaches, this metric is computationally efficient, interpretable, and adaptable to various use cases, offering a scalable and practical solution to assess context utilization in open-book QA systems.

Manuel Accettulli Huber, Adriana Correia, Sanjaye Ramgoolam et al.

The Linguistic Matrix Theory programme introduced by Kartsaklis, Ramgoolam and Sadrzadeh is an approach to the statistics of matrices that are generated in type-driven distributional semantics, based on permutation invariant polynomial functions which are regarded as the key observables encoding the significant statistics. In this paper we generalize the previous results on the approximate Gaussianity of matrix distributions arising from compositional distributional semantics. We also introduce a geometry of observable vectors for words, defined by exploiting the graph-theoretic basis for the permutation invariants and the statistical characteristics of the ensemble of matrices associated with the words. We describe successful applications of this unified framework to a number of tasks in computational linguistics, associated with the distinctions between synonyms, antonyms, hypernyms and hyponyms.