Existing Reward Models (RMs), typically trained on general preference data, struggle in Retrieval Augmented Generation (RAG) settings, which require judging responses for faithfulness to retrieved context, relevance to the user query, appropriate refusals when context is insufficient, completeness and conciseness of information. To address the lack of publicly available RAG-centric preference datasets and specialised RMs, we introduce RAGferee, a methodology that repurposes question-answering (QA) datasets into preference pairs that prioritise groundedness over stylistic features, enabling the training of contextual RMs better suited to judging RAG responses. Using RAGferee, we curate a small preference dataset of 4K samples and fine-tune RMs ranging from 7B to 24B parameters. Our RAG-centric RMs achieve state-of-the-art performance on ContextualJudgeBench, surpassing existing 70B+ RMs trained on much larger (up to 2.4M samples) general corpora, with an absolute improvement of +15.5%.

We present GaRAGe, a large RAG benchmark with human-curated long-form answers and annotations of each grounding passage, allowing a fine-grained evaluation of whether LLMs can identify relevant grounding when generating RAG answers. Our benchmark contains 2366 questions of diverse complexity, dynamism, and topics, and includes over 35K annotated passages retrieved from both private document sets and the Web, to reflect real-world RAG use cases. This makes it an ideal test bed to evaluate an LLM’s ability to identify only the relevant information necessary to compose a response, or provide a deflective response when there is insufficient information. Evaluations of multiple state-of-the-art LLMs on GaRAGe show that the models tend to over-summarise rather than (a) ground their answers strictly on the annotated relevant passages (reaching at most a Relevance-Aware Factuality Score of 60%), or (b) deflect when no relevant grounding is available (reaching at most 31% true positive rate in deflections). The F₁ in attribution to relevant sources is at most 58.9%, and we show that performance is particularly reduced when answering time-sensitive questions and when having to draw knowledge from sparser private grounding sources.

Long-form question answering (LFQA) aims at generating in-depth answers to end-user questions, providing relevant information beyond the direct answer. However, existing retrievers are typically optimized towards information that directly targets the question, missing out on such contextual information. Furthermore, there is a lack of training data for relevant context. To this end, we propose and compare different weak supervision techniques to optimize retrieval for contextual information. Experiments demonstrate improvements on the end-to-end QA performance on ASQA, a dataset for long-form question answering. Importantly, as more contextual information is retrieved, we improve the relevant page recall for LFQA by 14.7% and the groundedness of generated long-form answers by 12.5%. Finally, we show that long-form answers often anticipate likely follow-up questions, via experiments on a conversational QA dataset.

To successfully account for language, computational models need to take into account both the linguistic context (the content of the utterances) and the extra-linguistic context (for instance, the participants in a dialogue). We focus on a referential task that asks models to link entity mentions in a TV show to the corresponding characters, and design an architecture that attempts to account for both kinds of context. In particular, our architecture combines a previously proposed specialized module (an “entity library”) for character representation with transfer learning from a pre-trained language model. We find that, although the model does improve linguistic contextualization, it fails to successfully integrate extra-linguistic information about the participants in the dialogue. Our work shows that it is very challenging to incorporate extra-linguistic information into pre-trained language models.

In recent years, the NLP community has shown increasing interest in analysing how deep learning models work. Given that large models trained on complex tasks are difficult to inspect, some of this work has focused on controlled tasks that emulate specific aspects of language. We propose a new set of such controlled tasks to explore a crucial aspect of natural language processing that has not received enough attention: the need to retrieve discrete information from sequences. We also study model behavior on the tasks with simple instantiations of Transformers and LSTMs. Our results highlight the beneficial role of decoder attention and its sometimes unexpected interaction with other components. Moreover, we show that, for most of the tasks, these simple models still show significant difficulties. We hope that the community will take up the analysis possibilities that our tasks afford, and that a clearer understanding of model behavior on the tasks will lead to better and more transparent models.

Language models keep track of complex information about the preceding context – including, e.g., syntactic relations in a sentence. We investigate whether they also capture information beneficial for resolving pronominal anaphora in English. We analyze two state of the art models with LSTM and Transformer architectures, via probe tasks and analysis on a coreference annotated corpus. The Transformer outperforms the LSTM in all analyses. Our results suggest that language models are more successful at learning grammatical constraints than they are at learning truly referential information, in the sense of capturing the fact that we use language to refer to entities in the world. However, we find traces of the latter aspect, too.

Humans use language to refer to entities in the external world. Motivated by this, in recent years several models that incorporate a bias towards learning entity representations have been proposed. Such entity-centric models have shown empirical success, but we still know little about why. In this paper we analyze the behavior of two recently proposed entity-centric models in a referential task, Entity Linking in Multi-party Dialogue (SemEval 2018 Task 4). We show that these models outperform the state of the art on this task, and that they do better on lower frequency entities than a counterpart model that is not entity-centric, with the same model size. We argue that making models entity-centric naturally fosters good architectural decisions. However, we also show that these models do not really build entity representations and that they make poor use of linguistic context. These negative results underscore the need for model analysis, to test whether the motivations for particular architectures are borne out in how models behave when deployed.

The present work investigates whether different quantification mechanisms (set comparison, vague quantification, and proportional estimation) can be jointly learned from visual scenes by a multi-task computational model. The motivation is that, in humans, these processes underlie the same cognitive, non-symbolic ability, which allows an automatic estimation and comparison of set magnitudes. We show that when information about lower-complexity tasks is available, the higher-level proportional task becomes more accurate than when performed in isolation. Moreover, the multi-task model is able to generalize to unseen combinations of target/non-target objects. Consistently with behavioral evidence showing the interference of absolute number in the proportional task, the multi-task model no longer works when asked to provide the number of target objects in the scene.

This paper describes our winning contribution to SemEval 2018 Task 4: Character Identification on Multiparty Dialogues. It is a simple, standard model with one key innovation, an entity library. Our results show that this innovation greatly facilitates the identification of infrequent characters. Because of the generic nature of our model, this finding is potentially relevant to any task that requires the effective learning from sparse or imbalanced data.

Topics generated by topic models are typically presented as a list of topic terms. Automatic topic labelling is the task of generating a succinct label that summarises the theme or subject of a topic, with the intention of reducing the cognitive load of end-users when interpreting these topics. Traditionally, topic label systems focus on a single label modality, e.g. textual labels. In this work we propose a multimodal approach to topic labelling using a simple feedforward neural network. Given a topic and a candidate image or textual label, our method automatically generates a rating for the label, relative to the topic. Experiments show that this multimodal approach outperforms single-modality topic labelling systems.

In this article we propose a rank aggregation method for the task of collocations detection. It consists of applying some well-known methods (e.g. Dice method, chi-square test, z-test and likelihood ratio) and then aggregating the resulting collocations rankings by rank distance and Borda score. These two aggregation methods are especially well suited for the task, since the results of each individual method naturally forms a ranking of collocations. Combination methods are known to usually improve the results, and indeed, the proposed aggregation method performs better then each individual method taken in isolation.