Arash Eshghi

Javier Chiyah-Garcia, Alessandro Suglia, Arash Eshghi

In dialogue, the addressee may initially misunderstand the speaker and respond erroneously, often prompting the speaker to correct the misunderstanding in the next turn with a Third Position Repair (TPR). The ability to process and respond appropriately to such repair sequences is thus crucial in conversational AI systems. In this paper, we first collect, analyse, and publicly release BlockWorld-Repairs: a dataset of multi-modal TPR sequences in an instruction-following manipulation task that is, by design, rife with referential ambiguity. We employ this dataset to evaluate several state-of-the-art Vision and Language Models (VLM) across multiple settings, focusing on their capability to process and accurately respond to TPRs and thus recover from miscommunication. We find that, compared to humans, all models significantly underperform in this task. We then show that VLMs can benefit from specialised losses targeting relevant tokens during fine-tuning, achieving better performance and generalising better to new scenarios. Our results suggest that these models are not yet ready to be deployed in multi-modal collaborative settings where repairs are common, and highlight the need to design training regimes and objectives that facilitate learning from interaction. Our code and data are available at www.github.com/JChiyah/blockworld-repairs

Georgios Pantazopoulos, Malvina Nikandrou, Amit Parekh et al.

Interactive and embodied tasks pose at least two fundamental challenges to existing Vision & Language (VL) models, including 1) grounding language in trajectories of actions and observations, and 2) referential disambiguation. To tackle these challenges, we propose an Embodied MultiModal Agent (EMMA): a unified encoder-decoder model that reasons over images and trajectories, and casts action prediction as multimodal text generation. By unifying all tasks as text generation, EMMA learns a language of actions which facilitates transfer across tasks. Different to previous modular approaches with independently trained components, we use a single multitask model where each task contributes to goal completion. EMMA performs on par with similar models on several VL benchmarks and sets a new state-of-the-art performance (36.81% success rate) on the Dialog-guided Task Completion (DTC), a benchmark to evaluate dialog-guided agents in the Alexa Arena

Vevake Balaraman, Arash Eshghi, Ioannis Konstas et al.

The ability to handle miscommunication is crucial to robust and faithful conversational AI. People usually deal with miscommunication immediately as they detect it, using highly systematic interactional mechanisms called repair. One important type of repair is Third Position Repair (TPR) whereby a speaker is initially misunderstood but then corrects the misunderstanding as it becomes apparent after the addressee's erroneous response. Here, we collect and publicly release Repair-QA, the first large dataset of TPRs in a conversational question answering (QA) setting. The data is comprised of the TPR turns, corresponding dialogue contexts, and candidate repairs of the original turn for execution of TPRs. We demonstrate the usefulness of the data by training and evaluating strong baseline models for executing TPRs. For stand-alone TPR execution, we perform both automatic and human evaluations on a fine-tuned T5 model, as well as OpenAI's GPT-3 LLMs. Additionally, we extrinsically evaluate the LLMs' TPR processing capabilities in the downstream conversational QA task. The results indicate poor out-of-the-box performance on TPR's by the GPT-3 models, which then significantly improves when exposed to Repair-QA.

Georgios Pantazopoulos, Malvina Nikandrou, Alessandro Suglia et al.

This study explores replacing Transformers in Visual Language Models (VLMs) with Mamba, a recent structured state space model (SSM) that demonstrates promising performance in sequence modeling. We test models up to 3B parameters under controlled conditions, showing that Mamba-based VLMs outperforms Transformers-based VLMs in captioning, question answering, and reading comprehension. However, we find that Transformers achieve greater performance in visual grounding and the performance gap widens with scale. We explore two hypotheses to explain this phenomenon: 1) the effect of task-agnostic visual encoding on the updates of the hidden states, and 2) the difficulty in performing visual grounding from the perspective of in-context multimodal retrieval. Our results indicate that a task-aware encoding yields minimal performance gains on grounding, however, Transformers significantly outperform Mamba at in-context multimodal retrieval. Overall, Mamba shows promising performance on tasks where the correct output relies on a summary of the image but struggles when retrieval of explicit information from the context is required.

Javier Chiyah-Garcia, Alessandro Suglia, Arash Eshghi et al.

Referential ambiguities arise in dialogue when a referring expression does not uniquely identify the intended referent for the addressee. Addressees usually detect such ambiguities immediately and work with the speaker to repair it using meta-communicative, Clarificational Exchanges (CE): a Clarification Request (CR) and a response. Here, we argue that the ability to generate and respond to CRs imposes specific constraints on the architecture and objective functions of multi-modal, visually grounded dialogue models. We use the SIMMC 2.0 dataset to evaluate the ability of different state-of-the-art model architectures to process CEs, with a metric that probes the contextual updates that arise from them in the model. We find that language-based models are able to encode simple multi-modal semantic information and process some CEs, excelling with those related to the dialogue history, whilst multi-modal models can use additional learning objectives to obtain disentangled object representations, which become crucial to handle complex referential ambiguities across modalities overall.

Arash Eshghi, Arash Ashrafzadeh

In conversation, speakers produce language incrementally, word by word, while continuously monitoring the appropriateness of their own contribution in the dynamically unfolding context of the conversation; and this often leads them to repair their own utterance on the fly. This real-time language processing capacity is furthermore crucial to the development of fluent and natural conversational AI. In this paper, we use a previously learned Dynamic Syntax grammar and the CHILDES corpus to develop, train and evaluate a probabilistic model for incremental generation where input to the model is a purely semantic generation goal concept in Type Theory with Records (TTR). We show that the model's output exactly matches the gold candidate in 78% of cases with a ROUGE-l score of 0.86. We further do a zero-shot evaluation of the ability of the same model to generate self-repairs when the generation goal changes mid-utterance. Automatic evaluation shows that the model can generate self-repairs correctly in 85% of cases. A small human evaluation confirms the naturalness and grammaticality of the generated self-repairs. Overall, these results further highlight the generalisation power of grammar-based models and lay the foundations for more controllable, and naturally interactive conversational AI systems.

Alessandro Suglia, Claudio Greco, Katie Baker et al.

AI personal assistants deployed via robots or wearables require embodied understanding to collaborate with humans effectively. However, current Vision-Language Models (VLMs) primarily focus on third-person view videos, neglecting the richness of egocentric perceptual experience. To address this gap, we propose three key contributions. First, we introduce the Egocentric Video Understanding Dataset (EVUD) for training VLMs on video captioning and question answering tasks specific to egocentric videos. Second, we present AlanaVLM, a 7B parameter VLM trained using parameter-efficient methods on EVUD. Finally, we evaluate AlanaVLM's capabilities on OpenEQA, a challenging benchmark for embodied video question answering. Our model achieves state-of-the-art performance, outperforming open-source models including strong Socratic models using GPT-4 as a planner by 3.6%. Additionally, we outperform Claude 3 and Gemini Pro Vision 1.0 and showcase competitive results compared to Gemini Pro 1.5 and GPT-4V, even surpassing the latter in spatial reasoning. This research paves the way for building efficient VLMs that can be deployed in robots or wearables, leveraging embodied video understanding to collaborate seamlessly with humans in everyday tasks, contributing to the next generation of Embodied AI.

Georgios Pantazopoulos, Alessandro Suglia, Oliver Lemon et al.

An effective method for combining frozen large language models (LLM) and visual encoders involves a resampler module that creates a `visual prompt' which is provided to the LLM, along with the textual prompt. While this approach has enabled impressive performance across many coarse-grained tasks like image captioning and visual question answering, more fine-grained tasks that require spatial understanding have not been thoroughly examined. In this paper, we use \textit{diagnostic classifiers} to measure the extent to which the visual prompt produced by the resampler encodes spatial information. Our results show that this information is largely absent from the resampler output when kept frozen during training of the classifiers. However, when the resampler and classifier are trained jointly, we observe a significant performance boost. This shows that the compression achieved by the resamplers can in principle encode the requisite spatial information, but that more object-aware objectives are needed at the pretraining stage to facilitate this capability

Houman Mehrafarin, Arash Eshghi, Ioannis Konstas

Evaluating Large Language Models (LLMs) on reasoning benchmarks demonstrates their ability to solve compositional questions. However, little is known of whether these models engage in genuine logical reasoning or simply rely on implicit cues to generate answers. In this paper, we investigate the transitive reasoning capabilities of two distinct LLM architectures, LLaMA 2 and Flan-T5, by manipulating facts within two compositional datasets: QASC and Bamboogle. We controlled for potential cues that might influence the models' performance, including (a) word/phrase overlaps across sections of test input; (b) models' inherent knowledge during pre-training or fine-tuning; and (c) Named Entities. Our findings reveal that while both models leverage (a), Flan-T5 shows more resilience to experiments (b and c), having less variance than LLaMA 2. This suggests that models may develop an understanding of transitivity through fine-tuning on knowingly relevant datasets, a hypothesis we leave to future work.

Sabrina Chiesurin, Dimitris Dimakopoulos, Marco Antonio Sobrevilla Cabezudo et al.

Large language models are known to produce output which sounds fluent and convincing, but is also often wrong, e.g. "unfaithful" with respect to a rationale as retrieved from a knowledge base. In this paper, we show that task-based systems which exhibit certain advanced linguistic dialog behaviors, such as lexical alignment (repeating what the user said), are in fact preferred and trusted more, whereas other phenomena, such as pronouns and ellipsis are dis-preferred. We use open-domain question answering systems as our test-bed for task based dialog generation and compare several open- and closed-book models. Our results highlight the danger of systems that appear to be trustworthy by parroting user input while providing an unfaithful response.

Javier Chiyah-Garcia, Alessandro Suglia, José Lopes et al.

Anaphoric expressions, such as pronouns and referential descriptions, are situated with respect to the linguistic context of prior turns, as well as, the immediate visual environment. However, a speaker's referential descriptions do not always uniquely identify the referent, leading to ambiguities in need of resolution through subsequent clarificational exchanges. Thus, effective Ambiguity Detection and Coreference Resolution are key to task success in Conversational AI. In this paper, we present models for these two tasks as part of the SIMMC 2.0 Challenge (Kottur et al. 2021). Specifically, we use TOD-BERT and LXMERT based models, compare them to a number of baselines and provide ablation experiments. Our results show that (1) language models are able to exploit correlations in the data to detect ambiguity; and (2) unimodal coreference resolution models can avoid the need for a vision component, through the use of smart object representations.

Miruna Clinciu, Arash Eshghi, Helen Hastie

As transparency becomes key for robotics and AI, it will be necessary to evaluate the methods through which transparency is provided, including automatically generated natural language (NL) explanations. Here, we explore parallels between the generation of such explanations and the much-studied field of evaluation of Natural Language Generation (NLG). Specifically, we investigate which of the NLG evaluation measures map well to explanations. We present the ExBAN corpus: a crowd-sourced corpus of NL explanations for Bayesian Networks. We run correlations comparing human subjective ratings with NLG automatic measures. We find that embedding-based automatic NLG evaluation methods, such as BERTScore and BLEURT, have a higher correlation with human ratings, compared to word-overlap metrics, such as BLEU and ROUGE. This work has implications for Explainable AI and transparent robotic and autonomous systems.

Igor Shalyminov, Sungjin Lee, Arash Eshghi et al.

Goal-oriented dialogue systems are now being widely adopted in industry where it is of key importance to maintain a rapid prototyping cycle for new products and domains. Data-driven dialogue system development has to be adapted to meet this requirement --- therefore, reducing the amount of data and annotations necessary for training such systems is a central research problem. In this paper, we present the Dialogue Knowledge Transfer Network (DiKTNet), a state-of-the-art approach to goal-oriented dialogue generation which only uses a few example dialogues (i.e. few-shot learning), none of which has to be annotated. We achieve this by performing a 2-stage training. Firstly, we perform unsupervised dialogue representation pre-training on a large source of goal-oriented dialogues in multiple domains, the MetaLWOz corpus. Secondly, at the transfer stage, we train DiKTNet using this representation together with 2 other textual knowledge sources with different levels of generality: ELMo encoder and the main dataset's source domains. Our main dataset is the Stanford Multi-Domain dialogue corpus. We evaluate our model on it in terms of BLEU and Entity F1 scores, and show that our approach significantly and consistently improves upon a series of baseline models as well as over the previous state-of-the-art dialogue generation model, ZSDG. The improvement upon the latter --- up to 10% in Entity F1 and the average of 3% in BLEU score --- is achieved using only the equivalent of 10% of ZSDG's in-domain training data.

Angus Addlesee, Arash Eshghi, Ioannis Konstas

Dialogue technologies such as Amazon's Alexa have the potential to transform the healthcare industry. However, current systems are not yet naturally interactive: they are often turn-based, have naive end-of-turn detection and completely ignore many types of verbal and visual feedback - such as backchannels, hesitation markers, filled pauses, gaze, brow furrows and disfluencies - that are crucial in guiding and managing the conversational process. This is especially important in the healthcare industry as target users of Spoken Dialogue Systems (SDSs) are likely to be frail, older, distracted or suffer from cognitive decline which impacts their ability to make effective use of current systems. In this paper, we outline some of the challenges that are in urgent need of further research, including Incremental Speech Recognition and a systematic study of the interactional patterns in conversation that are potentially diagnostic of dementia, and how these might inform research on and the design of the next generation of SDSs.

Igor Shalyminov, Sungjin Lee, Arash Eshghi et al.

Learning with minimal data is one of the key challenges in the development of practical, production-ready goal-oriented dialogue systems. In a real-world enterprise setting where dialogue systems are developed rapidly and are expected to work robustly for an ever-growing variety of domains, products, and scenarios, efficient learning from a limited number of examples becomes indispensable. In this paper, we introduce a technique to achieve state-of-the-art dialogue generation performance in a few-shot setup, without using any annotated data. We do this by leveraging background knowledge from a larger, more highly represented dialogue source --- namely, the MetaLWOz dataset. We evaluate our model on the Stanford Multi-Domain Dialogue Dataset, consisting of human-human goal-oriented dialogues in in-car navigation, appointment scheduling, and weather information domains. We show that our few-shot approach achieves state-of-the art results on that dataset by consistently outperforming the previous best model in terms of BLEU and Entity F1 scores, while being more data-efficient by not requiring any data annotation.

Xingkun Liu, Arash Eshghi, Pawel Swietojanski et al.

We have recently seen the emergence of several publicly available Natural Language Understanding (NLU) toolkits, which map user utterances to structured, but more abstract, Dialogue Act (DA) or Intent specifications, while making this process accessible to the lay developer. In this paper, we present the first wide coverage evaluation and comparison of some of the most popular NLU services, on a large, multi-domain (21 domains) dataset of 25K user utterances that we have collected and annotated with Intent and Entity Type specifications and which will be released as part of this submission. The results show that on Intent classification Watson significantly outperforms the other platforms, namely, Dialogflow, LUIS and Rasa; though these also perform well. Interestingly, on Entity Type recognition, Watson performs significantly worse due to its low Precision. Again, Dialogflow, LUIS and Rasa perform well on this task.

Igor Shalyminov, Arash Eshghi, Oliver Lemon

Spontaneous spoken dialogue is often disfluent, containing pauses, hesitations, self-corrections and false starts. Processing such phenomena is essential in understanding a speaker's intended meaning and controlling the flow of the conversation. Furthermore, this processing needs to be word-by-word incremental to allow further downstream processing to begin as early as possible in order to handle real spontaneous human conversational behaviour. In addition, from a developer's point of view, it is highly desirable to be able to develop systems which can be trained from `clean' examples while also able to generalise to the very diverse disfluent variations on the same data -- thereby enhancing both data-efficiency and robustness. In this paper, we present a multi-task LSTM-based model for incremental detection of disfluency structure, which can be hooked up to any component for incremental interpretation (e.g. an incremental semantic parser), or else simply used to `clean up' the current utterance as it is being produced. We train the system on the Switchboard Dialogue Acts (SWDA) corpus and present its accuracy on this dataset. Our model outperforms prior neural network-based incremental approaches by about 10 percentage points on SWDA while employing a simpler architecture. To test the model's generalisation potential, we evaluate the same model on the bAbI+ dataset, without any additional training. bAbI+ is a dataset of synthesised goal-oriented dialogues where we control the distribution of disfluencies and their types. This shows that our approach has good generalisation potential, and sheds more light on which types of disfluency might be amenable to domain-general processing.

Yanchao Yu, Arash Eshghi, Gregory Mills et al.

We motivate and describe a new freely available human-human dialogue dataset for interactive learning of visually grounded word meanings through ostensive definition by a tutor to a learner. The data has been collected using a novel, character-by-character variant of the DiET chat tool (Healey et al., 2003; Mills and Healey, submitted) with a novel task, where a Learner needs to learn invented visual attribute words (such as " burchak " for square) from a tutor. As such, the text-based interactions closely resemble face-to-face conversation and thus contain many of the linguistic phenomena encountered in natural, spontaneous dialogue. These include self-and other-correction, mid-sentence continuations, interruptions, overlaps, fillers, and hedges. We also present a generic n-gram framework for building user (i.e. tutor) simulations from this type of incremental data, which is freely available to researchers. We show that the simulations produce outputs that are similar to the original data (e.g. 78% turn match similarity). Finally, we train and evaluate a Reinforcement Learning dialogue control agent for learning visually grounded word meanings, trained from the BURCHAK corpus. The learned policy shows comparable performance to a rule-based system built previously.

Yanchao Yu, Arash Eshghi, Oliver Lemon

We present a multi-modal dialogue system for interactive learning of perceptually grounded word meanings from a human tutor. The system integrates an incremental, semantic parsing/generation framework - Dynamic Syntax and Type Theory with Records (DS-TTR) - with a set of visual classifiers that are learned throughout the interaction and which ground the meaning representations that it produces. We use this system in interaction with a simulated human tutor to study the effects of different dialogue policies and capabilities on the accuracy of learned meanings, learning rates, and efforts/costs to the tutor. We show that the overall performance of the learning agent is affected by (1) who takes initiative in the dialogues; (2) the ability to express/use their confidence level about visual attributes; and (3) the ability to process elliptical and incrementally constructed dialogue turns. Ultimately, we train an adaptive dialogue policy which optimises the trade-off between classifier accuracy and tutoring costs.

Yanchao Yu, Arash Eshghi, Oliver Lemon

We present an optimised multi-modal dialogue agent for interactive learning of visually grounded word meanings from a human tutor, trained on real human-human tutoring data. Within a life-long interactive learning period, the agent, trained using Reinforcement Learning (RL), must be able to handle natural conversations with human users and achieve good learning performance (accuracy) while minimising human effort in the learning process. We train and evaluate this system in interaction with a simulated human tutor, which is built on the BURCHAK corpus -- a Human-Human Dialogue dataset for the visual learning task. The results show that: 1) The learned policy can coherently interact with the simulated user to achieve the goal of the task (i.e. learning visual attributes of objects, e.g. colour and shape); and 2) it finds a better trade-off between classifier accuracy and tutoring costs than hand-crafted rule-based policies, including ones with dynamic policies.

Arash Eshghi, Igor Shalyminov, Oliver Lemon

We investigate an end-to-end method for automatically inducing task-based dialogue systems from small amounts of unannotated dialogue data. It combines an incremental semantic grammar - Dynamic Syntax and Type Theory with Records (DS-TTR) - with Reinforcement Learning (RL), where language generation and dialogue management are a joint decision problem. The systems thus produced are incremental: dialogues are processed word-by-word, shown previously to be essential in supporting natural, spontaneous dialogue. We hypothesised that the rich linguistic knowledge within the grammar should enable a combinatorially large number of dialogue variations to be processed, even when trained on very few dialogues. Our experiments show that our model can process 74% of the Facebook AI bAbI dataset even when trained on only 0.13% of the data (5 dialogues). It can in addition process 65% of bAbI+, a corpus we created by systematically adding incremental dialogue phenomena such as restarts and self-corrections to bAbI. We compare our model with a state-of-the-art retrieval model, MemN2N. We find that, in terms of semantic accuracy, MemN2N shows very poor robustness to the bAbI+ transformations even when trained on the full bAbI dataset.

Igor Shalyminov, Arash Eshghi, Oliver Lemon

Natural, spontaneous dialogue proceeds incrementally on a word-by-word basis; and it contains many sorts of disfluency such as mid-utterance/sentence hesitations, interruptions, and self-corrections. But training data for machine learning approaches to dialogue processing is often either cleaned-up or wholly synthetic in order to avoid such phenomena. The question then arises of how well systems trained on such clean data generalise to real spontaneous dialogue, or indeed whether they are trainable at all on naturally occurring dialogue data. To answer this question, we created a new corpus called bAbI+ by systematically adding natural spontaneous incremental dialogue phenomena such as restarts and self-corrections to the Facebook AI Research's bAbI dialogues dataset. We then explore the performance of a state-of-the-art retrieval model, MemN2N, on this more natural dataset. Results show that the semantic accuracy of the MemN2N model drops drastically; and that although it is in principle able to learn to process the constructions in bAbI+, it needs an impractical amount of training data to do so. Finally, we go on to show that an incremental, semantic parser -- DyLan -- shows 100% semantic accuracy on both bAbI and bAbI+, highlighting the generalisation properties of linguistically informed dialogue models.

Dimitrios Kalatzis, Arash Eshghi, Oliver Lemon

We present a method for inducing new dialogue systems from very small amounts of unannotated dialogue data, showing how word-level exploration using Reinforcement Learning (RL), combined with an incremental and semantic grammar - Dynamic Syntax (DS) - allows systems to discover, generate, and understand many new dialogue variants. The method avoids the use of expensive and time-consuming dialogue act annotations, and supports more natural (incremental) dialogues than turn-based systems. Here, language generation and dialogue management are treated as a joint decision/optimisation problem, and the MDP model for RL is constructed automatically. With an implemented system, we show that this method enables a wide range of dialogue variations to be automatically captured, even when the system is trained from only a single dialogue. The variants include question-answer pairs, over- and under-answering, self- and other-corrections, clarification interaction, split-utterances, and ellipsis. This generalisation property results from the structural knowledge and constraints present within the DS grammar, and highlights some limitations of recent systems built using machine learning techniques only.