The usual way to interpret language models (LMs) is to test their performance on different benchmarks and subsequently infer their internal processes.In this paper, we present an alternative approach, concentrating on the _quality_ of LM processing, with a focus on their language abilities.To this end, we construct ‘linguistic task spaces’ – representations of an LM’s language conceptualisation – that shed light on the connections LMs draw between language phenomena.Task spaces are based on the interactions of the learning signals from different linguistic phenomena, which we assess via a method we call ‘similarity probing’.To disentangle the learning signals of linguistic phenomena, we further introduce a method called ‘fine-tuning via gradient differentials’ (FTGD).We apply our methods to language models of three different scales and find that larger models generalise better to overarching general concepts for linguistic tasks, making better use of their shared structure. Further, the distributedness of linguistic processing increases with pre-training through increased parameter sharing between related linguistic tasks. The overall generalisation patterns are mostly stable throughout training and not marked by incisive stages, potentially explaining the lack of successful curriculum strategies for LMs.

We study the communication of concepts at different levels of abstraction and in different contexts in an agent-based, interactive reference game. While playing the concept-level reference game, the neural network agents develop a communication system from scratch. We use a novel symbolic dataset that disentangles concept type (ranging from specific to generic) and context (ranging from fine to coarse) to study the influence of these factors on the emerging language. We compare two game scenarios: one in which speaker agents have access to context information (context-aware) and one in which the speaker agents do not have access to context information (context-unaware). First, we find that the agents learn higher-level concepts from the object inputs alone. Second, an analysis of the emergent communication system shows that only context-aware agents learn to communicate efficiently by adapting their messages to the context conditions and relying on context for unambiguous reference. Crucially, this behavior is not explicitly incentivized by the game, but efficient communication emerges and is driven by the availability of context alone. The emerging language we observe is reminiscent of evolutionary pressures on human languages and highlights the pivotal role of context in a communication system.

Finding the best way of adapting pre-trained language models to a task is a big challenge in current NLP. Just like the previous generation of task-tuned models (TT), models that are adapted to tasks via in-context-learning (ICL) or instruction tuning (IT) are robust in some setups, but not in others. Here, we present a detailed analysis of which design choices cause instabilities and inconsistencies in LLM predictions. First, we show how spurious correlations between input distributions and labels – a known issue in TT models – form only a minor problem for prompted models. Then we engage in a systematic, holistic evaluation of different factors that have been found to influence predictions in a prompting setup. We test all possible combinations of a range of factors on both vanilla and instruction-tuned LLMs of different scale, and statistically analyse the results to show which factors are the most influential, the most interactive or the most stable. From our results, we deduce which factors can be used without precautions, should be avoided or handled with care in most settings.

At the staggering pace with which the capabilities of large language models (LLMs) are increasing, creating future-proof evaluation sets to assess their understanding becomes more and more challenging. In this paper, we propose a novel paradigm for evaluating LLMs which leverages the idea that correct world understanding should be consistent across different (Fregean) senses of the same meaning. Accordingly, we measure understanding not in terms of correctness but by evaluating consistency across multiple senses that are generated by the model itself. We showcase our approach by instantiating a test where the different senses are different languages, hence using multilingual self-consistency as a litmus test for the model’s understanding and simultaneously addressing the important topic of multilingualism. Taking one of the latest versions of ChatGPT as our object of study, we evaluate multilingual consistency for two different tasks across three different languages. We show that its multilingual consistency is still lacking, and that its task and world understanding are thus not language-independent. As our approach does not require any static evaluation corpora in languages other than English, it can easily and cheaply be extended to different languages and tasks and could become an integral part of future benchmarking efforts.

Obtaining human-like performance in NLP is often argued to require compositional generalisation. Whether neural networks exhibit this ability is usually studied by training models on highly compositional synthetic data. However, compositionality in natural language is much more complex than the rigid, arithmetic-like version such data adheres to, and artificial compositionality tests thus do not allow us to determine how neural models deal with more realistic forms of compositionality. In this work, we re-instantiate three compositionality tests from the literature and reformulate them for neural machine translation (NMT).Our results highlight that: i) unfavourably, models trained on more data are more compositional; ii) models are sometimes less compositional than expected, but sometimes more, exemplifying that different levels of compositionality are required, and models are not always able to modulate between them correctly; iii) some of the non-compositional behaviours are mistakes, whereas others reflect the natural variation in data. Apart from an empirical study, our work is a call to action: we should rethink the evaluation of compositionality in neural networks and develop benchmarks using real data to evaluate compositionality on natural language, where composing meaning is not as straightforward as doing the math.

In natural language, referencing objects at different levels of specificity is a fundamental pragmatic mechanism for efficient communication in context. We develop a novel communication game, the hierarchical reference game, to study the emergence of such reference systems in artificial agents. We consider a simplified world, in which concepts are abstractions over a set of primitive attributes (e.g., color, style, shape). Depending on how many attributes are combined, concepts are more general (“circle”) or more specific (“red dotted circle”). Based on the context, the agents have to communicate at different levels of this hierarchy. Our results show that the agents learn to play the game successfully and can even generalize to novel concepts. To achieve abstraction, they use implicit (omitting irrelevant information) and explicit (indicating that attributes are irrelevant) strategies. In addition, the compositional structure underlying the concept hierarchy is reflected in the emergent protocols, indicating that the need to develop hierarchical reference systems supports the emergence of compositionality.

In this paper, we propose to study language modelling as a multi-task problem, bringing together three strands of research: multi-task learning, linguistics, and interpretability. Based on hypotheses derived from linguistic theory, we investigate whether language models adhere to learning principles of multi-task learning during training. To showcase the idea, we analyse the generalisation behaviour of language models as they learn the linguistic concept of Negative Polarity Items (NPIs). Our experiments demonstrate that a multi-task setting naturally emerges within the objective of the more general task of language modelling. We argue that this insight is valuable for multi-task learning, linguistics and interpretability research and can lead to exciting new findings in all three domains.

Referential games offer a grounded learning environment for neural agents which accounts for the fact that language is functionally used to communicate. However, they do not take into account a second constraint considered to be fundamental for the shape of human language: that it must be learnable by new language learners. Cogswell et al. (2019) introduced cultural transmission within referential games through a changing population of agents to constrain the emerging language to be learnable. However, the resulting languages remain inherently biased by the agents’ underlying capabilities. In this work, we introduce Language Transmission Simulator to model both cultural and architectural evolution in a population of agents. As our core contribution, we empirically show that the optimal situation is to take into account also the learning biases of the language learners and thus let language and agents co-evolve. When we allow the agent population to evolve through architectural evolution, we achieve across the board improvements on all considered metrics and surpass the gains made with cultural transmission. These results stress the importance of studying the underlying agent architecture and pave the way to investigate the co-evolution of language and agent in language emergence studies.

Neural networks are surprisingly good at interpolating and perform remarkably well when the training set examples resemble those in the test set. However, they are often unable to extrapolate patterns beyond the seen data, even when the abstractions required for such patterns are simple. In this paper, we first review the notion of extrapolation, why it is important and how one could hope to tackle it. We then focus on a specific type of extrapolation which is especially useful for natural language processing: generalization to sequences that are longer than the training ones. We hypothesize that models with a separate content- and location-based attention are more likely to extrapolate than those with common attention mechanisms. We empirically support our claim for recurrent seq2seq models with our proposed attention on variants of the Lookup Table task. This sheds light on some striking failures of neural models for sequences and on possible methods to approaching such issues.

In previous work, artificial agents were shown to achieve almost perfect accuracy in referential games where they have to communicate to identify images. Nevertheless, the resulting communication protocols rarely display salient features of natural languages, such as compositionality. In this paper, we propose some realistic sources of pressure on communication that avert this outcome. More specifically, we formalise the principle of least effort through an auxiliary objective. Moreover, we explore several game variants, inspired by the principle of object constancy, in which we alter the frequency, position, and luminosity of the objects in the images. We perform an extensive analysis on their effect through compositionality metrics, diagnostic classifiers, and zero-shot evaluation. Our findings reveal that the proposed sources of pressure result in emerging languages with less redundancy, more focus on high-level conceptual information, and better abilities of generalisation. Overall, our contributions reduce the gap between emergent and natural languages.

In this paper, we consider the syntactic properties of languages emerged in referential games, using unsupervised grammar induction (UGI) techniques originally designed to analyse natural language. We show that the considered UGI techniques are appropriate to analyse emergent languages and we then study if the languages that emerge in a typical referential game setup exhibit syntactic structure, and to what extent this depends on the maximum message length and number of symbols that the agents are allowed to use. Our experiments demonstrate that a certain message length and vocabulary size are required for structure to emerge, but they also illustrate that more sophisticated game scenarios are required to obtain syntactic properties more akin to those observed in human language. We argue that UGI techniques should be part of the standard toolkit for analysing emergent languages and release a comprehensive library to facilitate such analysis for future researchers.

This paper introduces the PhotoBook dataset, a large-scale collection of visually-grounded, task-oriented dialogues in English designed to investigate shared dialogue history accumulating during conversation. Taking inspiration from seminal work on dialogue analysis, we propose a data-collection task formulated as a collaborative game prompting two online participants to refer to images utilising both their visual context as well as previously established referring expressions. We provide a detailed description of the task setup and a thorough analysis of the 2,500 dialogues collected. To further illustrate the novel features of the dataset, we propose a baseline model for reference resolution which uses a simple method to take into account shared information accumulated in a reference chain. Our results show that this information is particularly important to resolve later descriptions and underline the need to develop more sophisticated models of common ground in dialogue interaction.

We propose a grounded dialogue state encoder which addresses a foundational issue on how to integrate visual grounding with dialogue system components. As a test-bed, we focus on the GuessWhat?! game, a two-player game where the goal is to identify an object in a complex visual scene by asking a sequence of yes/no questions. Our visually-grounded encoder leverages synergies between guessing and asking questions, as it is trained jointly using multi-task learning. We further enrich our model via a cooperative learning regime. We show that the introduction of both the joint architecture and cooperative learning lead to accuracy improvements over the baseline system. We compare our approach to an alternative system which extends the baseline with reinforcement learning. Our in-depth analysis shows that the linguistic skills of the two models differ dramatically, despite approaching comparable performance levels. This points at the importance of analyzing the linguistic output of competing systems beyond numeric comparison solely based on task success.

Previous research into agent communication has shown that a pre-trained guide can speed up the learning process of an imitation learning agent. The guide achieves this by providing the agent with discrete messages in an emerged language about how to solve the task. We extend this one-directional communication by a one-bit communication channel from the learner back to the guide: It is able to ask the guide for help, and we limit the guidance by penalizing the learner for these requests. During training, the agent learns to control this gate based on its current observation. We find that the amount of requested guidance decreases over time and guidance is requested in situations of high uncertainty. We investigate the agent’s performance in cases of open and closed gates and discuss potential motives for the observed gating behavior.

Learning to follow human instructions is a long-pursued goal in artificial intelligence. The task becomes particularly challenging if no prior knowledge of the employed language is assumed while relying only on a handful of examples to learn from. Work in the past has relied on hand-coded components or manually engineered features to provide strong inductive biases that make learning in such situations possible. In contrast, here we seek to establish whether this knowledge can be acquired automatically by a neural network system through a two phase training procedure: A (slow) offline learning stage where the network learns about the general structure of the task and a (fast) online adaptation phase where the network learns the language of a new given speaker. Controlled experiments show that when the network is exposed to familiar instructions but containing novel words, the model adapts very efficiently to the new vocabulary. Moreover, even for human speakers whose language usage can depart significantly from our artificial training language, our network can still make use of its automatically acquired inductive bias to learn to follow instructions more effectively.

Since their inception, encoder-decoder models have successfully been applied to a wide array of problems in computational linguistics. The most recent successes are predominantly due to the use of different variations of attention mechanisms, but their cognitive plausibility is questionable. In particular, because past representations can be revisited at any point in time, attention-centric methods seem to lack an incentive to build up incrementally more informative representations of incoming sentences. This way of processing stands in stark contrast with the way in which humans are believed to process language: continuously and rapidly integrating new information as it is encountered. In this work, we propose three novel metrics to assess the behavior of RNNs with and without an attention mechanism and identify key differences in the way the different model types process sentences.

While sequence-to-sequence models have shown remarkable generalization power across several natural language tasks, their construct of solutions are argued to be less compositional than human-like generalization. In this paper, we present seq2attn, a new architecture that is specifically designed to exploit attention to find compositional patterns in the input. In seq2attn, the two standard components of an encoder-decoder model are connected via a transcoder, that modulates the information flow between them. We show that seq2attn can successfully generalize, without requiring any additional supervision, on two tasks which are specifically constructed to challenge the compositional skills of neural networks. The solutions found by the model are highly interpretable, allowing easy analysis of both the types of solutions that are found and potential causes for mistakes. We exploit this opportunity to introduce a new paradigm to test compositionality that studies the extent to which a model overgeneralizes when confronted with exceptions. We show that seq2attn exhibits such overgeneralization to a larger degree than a standard sequence-to-sequence model.

We present a detailed comparison of two types of sequence to sequence models trained to conduct a compositional task. The models are architecturally identical at inference time, but differ in the way that they are trained: our baseline model is trained with a task-success signal only, while the other model receives additional supervision on its attention mechanism (Attentive Guidance), which has shown to be an effective method for encouraging more compositional solutions. We first confirm that the models with attentive guidance indeed infer more compositional solutions than the baseline, by training them on the lookup table task presented by Liska et al. (2019). We then do an in-depth analysis of the structural differences between the two model types, focusing in particular on the organisation of the parameter space and the hidden layer activations and find noticeable differences in both these aspects. Guided networks focus more on the components of the input rather than the sequence as a whole and develop small functional groups of neurons with specific purposes that use their gates more selectively. Results from parameter heat maps, component swapping and graph analysis also indicate that guided networks exhibit a more modular structure with a small number of specialized, strongly connected neurons.

Our goal is to explore how the abilities brought in by a dialogue manager can be included in end-to-end visually grounded conversational agents. We make initial steps towards this general goal by augmenting a task-oriented visual dialogue model with a decision-making component that decides whether to ask a follow-up question to identify a target referent in an image, or to stop the conversation to make a guess. Our analyses show that adding a decision making component produces dialogues that are less repetitive and that include fewer unnecessary questions, thus potentially leading to more efficient and less unnatural interactions.

We investigate the potential of adversarial evaluation methods for open-domain dialogue generation systems, comparing the performance of a discriminative agent to that of humans on the same task. Our results show that the task is hard, both for automated models and humans, but that a discriminative agent can learn patterns that lead to above-chance performance.