Oliver Lemon

Angus Addlesee, Weronika Sieińska, Nancie Gunson et al.

This paper evaluates the extent to which current Large Language Models (LLMs) can capture task-oriented multi-party conversations (MPCs). We have recorded and transcribed 29 MPCs between patients, their companions, and a social robot in a hospital. We then annotated this corpus for multi-party goal-tracking and intent-slot recognition. People share goals, answer each other's goals, and provide other people's goals in MPCs - none of which occur in dyadic interactions. To understand user goals in MPCs, we compared three methods in zero-shot and few-shot settings: we fine-tuned T5, created pre-training tasks to train DialogLM using LED, and employed prompt engineering techniques with GPT-3.5-turbo, to determine which approach can complete this novel task with limited data. GPT-3.5-turbo significantly outperformed the others in a few-shot setting. The `reasoning' style prompt, when given 7% of the corpus as example annotated conversations, was the best performing method. It correctly annotated 62.32% of the goal tracking MPCs, and 69.57% of the intent-slot recognition MPCs. A `story' style prompt increased model hallucination, which could be detrimental if deployed in safety-critical settings. We conclude that multi-party conversations still challenge state-of-the-art LLMs.

Laura Schauer, Jason Sweeney, Charlie Lyttle et al.

Today, conversational systems are expected to handle conversations in multi-party settings, especially within Socially Assistive Robots (SARs). However, practical usability remains difficult as there are additional challenges to overcome, such as speaker recognition, addressee recognition, and complex turn-taking. In this paper, we present our work on a multi-party conversational system, which invites two users to play a trivia quiz game. The system detects users' agreement or disagreement on a final answer and responds accordingly. Our evaluation includes both performance and user assessment results, with a focus on detecting user agreement. Our annotated transcripts and the code for the proposed system have been released open-source on GitHub.

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

Bhathiya Hemanthage, Christian Dondrup, Phil Bartie et al.

SimpleMTOD is a simple language model which recasts several sub-tasks in multimodal task-oriented dialogues as sequence prediction tasks. SimpleMTOD is built on a large-scale transformer-based auto-regressive architecture, which has already proven to be successful in uni-modal task-oriented dialogues, and effectively leverages transfer learning from pre-trained GPT-2. In-order to capture the semantics of visual scenes, we introduce both local and de-localized tokens for objects within a scene. De-localized tokens represent the type of an object rather than the specific object itself and so possess a consistent meaning across the dataset. SimpleMTOD achieves a state-of-the-art BLEU score (0.327) in the Response Generation sub-task of the SIMMC 2.0 test-std dataset while performing on par in other multimodal sub-tasks: Disambiguation, Coreference Resolution, and Dialog State Tracking. This is despite taking a minimalist approach for extracting visual (and non-visual) information. In addition the model does not rely on task-specific architectural changes such as classification heads.

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.

Angus Addlesee, Daniel Denley, Andy Edmondson et al.

Conversational agents participating in multi-party interactions face significant challenges in dialogue state tracking, since the identity of the speaker adds significant contextual meaning. It is common to utilise diarisation models to identify the speaker. However, it is not clear if these are accurate enough to correctly identify specific conversational events such as agreement or disagreement during a real-time interaction. This study uses a cooperative quiz, where the conversational agent acts as quiz-show host, to determine whether diarisation or a frequency-and-proximity-based method is more accurate at determining agreement, and whether this translates to feelings of engagement from the players. Experimental results show that our procedural system was more engaging to players, and was more accurate at detecting agreement, reaching an average accuracy of 0.44 compared to 0.28 for the diarised system.

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.

Marco Casadio, Tanvi Dinkar, Ekaterina Komendantskaya et al.

Machine Learning (ML) has exhibited substantial success in the field of Natural Language Processing (NLP). For example large language models have empirically proven to be capable of producing text of high complexity and cohesion. However, they are prone to inaccuracies and hallucinations. As these systems are increasingly integrated into real-world applications, ensuring their safety and reliability becomes a primary concern. There are safety critical contexts where such models must be robust to variability or attack, and give guarantees over their output. Computer Vision had pioneered the use of formal verification of neural networks for such scenarios and developed common verification standards and pipelines, leveraging precise formal reasoning about geometric properties of data manifolds. In contrast, NLP verification methods have only recently appeared in the literature. While presenting sophisticated algorithms, these papers have not yet crystallised into a common methodology. They are often light on the pragmatical issues of NLP verification and the area remains fragmented. In this paper, we attempt to distil and evaluate general components of an NLP verification pipeline, that emerges from the progress in the field to date. Our contributions are two-fold. Firstly, we propose a general methodology to analyse the effect of the embedding gap, a problem that refers to the discrepancy between verification of geometric subspaces and the semantic meaning of sentences, which the geometric subspaces are supposed to represent. We propose a number of practical NLP methods that can help to quantify the effects of the embedding gap. Secondly, we give a general method for training and verification of neural networks that leverages a more precise geometric estimation of semantic similarity of sentences in the embedding space and helps to overcome the effects of the embedding gap in practice.

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

Filippo Momentè, Alessandro Suglia, Mario Giulianelli et al.

We examine three evaluation paradigms: standard benchmarks (e.g., MMLU and BBH), interactive games (e.g., Signalling Games or Taboo), and cognitive tests (e.g., for working memory or theory of mind). First, we investigate which of the former two-benchmarks or games-is most effective at discriminating LLMs of varying quality. Then, inspired by human cognitive assessments, we compile a suite of targeted tests that measure cognitive abilities deemed essential for effective language use, and we investigate their correlation with model performance in benchmarks and games. Our analyses reveal that interactive games are superior to standard benchmarks in discriminating models. Causal and logical reasoning correlate with both static and interactive tests, while differences emerge regarding core executive functions and social/emotional skills, which correlate more with games. We advocate for the development of new interactive benchmarks and targeted cognitive tasks inspired by assessing human abilities but designed specifically for LLMs.

Xavier Alameda-Pineda, Angus Addlesee, Daniel Hernández García et al.

Despite the many recent achievements in developing and deploying social robotics, there are still many underexplored environments and applications for which systematic evaluation of such systems by end-users is necessary. While several robotic platforms have been used in gerontological healthcare, the question of whether or not a social interactive robot with multi-modal conversational capabilities will be useful and accepted in real-life facilities is yet to be answered. This paper is an attempt to partially answer this question, via two waves of experiments with patients and companions in a day-care gerontological facility in Paris with a full-sized humanoid robot endowed with social and conversational interaction capabilities. The software architecture, developed during the H2020 SPRING project, together with the experimental protocol, allowed us to evaluate the acceptability (AES) and usability (SUS) with more than 60 end-users. Overall, the users are receptive to this technology, especially when the robot perception and action skills are robust to environmental clutter and flexible to handle a plethora of different interactions.

Alessandro Suglia, Ioannis Konstas, Oliver Lemon

In recent years, several machine learning models have been proposed. They are trained with a language modelling objective on large-scale text-only data. With such pretraining, they can achieve impressive results on many Natural Language Understanding and Generation tasks. However, many facets of meaning cannot be learned by ``listening to the radio" only. In the literature, many Vision+Language (V+L) tasks have been defined with the aim of creating models that can ground symbols in the visual modality. In this work, we provide a systematic literature review of several tasks and models proposed in the V+L field. We rely on Wittgenstein's idea of `language games' to categorise such tasks into 3 different families: 1) discriminative games, 2) generative games, and 3) interactive games. Our analysis of the literature provides evidence that future work should be focusing on interactive games where communication in Natural Language is important to resolve ambiguities about object referents and action plans and that physical embodiment is essential to understand the semantics of situations and events. Overall, these represent key requirements for developing grounded meanings in neural models.

Nicola Horst, Davide Mazzaccara, Antonia Schmidt et al.

Interaction between learner and feedback-giver has come into focus recently for post-training of Large Language Models (LLMs), through the use of reward models that judge the appropriateness of a model's response. In this paper, we investigate whether Dialogue Games -- goal-directed and rule-governed activities driven predominantly by verbal actions -- can also serve as a source of feedback signals for learning. We introduce Playpen, an environment for off- and online learning through Dialogue Game self-play, and investigate a representative set of post-training methods: supervised fine-tuning; direct alignment (DPO); and reinforcement learning with GRPO. We experiment with post-training a small LLM (Llama-3.1-8B-Instruct), evaluating performance on unseen instances of training games as well as unseen games, and on standard benchmarks. We find that imitation learning through SFT improves performance on unseen instances, but negatively impacts other skills, while interactive learning with GRPO shows balanced improvements without loss of skills. We release the framework and the baseline training setups to foster research in the promising new direction of learning in (synthetic) interaction.

Hima Jacob Leven Suprabha, Laxmi Nag Laxminarayan Nagesh, Ajith Nair et al.

The integration of Large Language Models (LLMs) into multiagent systems has opened new possibilities for collaborative reasoning and cooperation with AI agents. This paper explores different prompting methods and evaluates their effectiveness in enhancing agent collaborative behaviour and decision-making. We enhance CoELA, a framework designed for building Collaborative Embodied Agents that leverage LLMs for multi-agent communication, reasoning, and task coordination in shared virtual spaces. Through systematic experimentation, we examine different LLMs and prompt engineering strategies to identify optimised combinations that maximise collaboration performance. Furthermore, we extend our research by integrating speech capabilities, enabling seamless collaborative voice-based interactions. Our findings highlight the effectiveness of prompt optimisation in enhancing collaborative agent performance; for example, our best combination improved the efficiency of the system running with Gemma3 by 22% compared to the original CoELA system. In addition, the speech integration provides a more engaging user interface for iterative system development and demonstrations.

Neeraj Cherakara, Finny Varghese, Sheena Shabana et al.

We demonstrate an embodied conversational agent that can function as a receptionist and generate a mixture of open and closed-domain dialogue along with facial expressions, by using a large language model (LLM) to develop an engaging conversation. We deployed the system onto a Furhat robot, which is highly expressive and capable of using both verbal and nonverbal cues during interaction. The system was designed specifically for the National Robotarium to interact with visitors through natural conversations, providing them with information about the facilities, research, news, upcoming events, etc. The system utilises the state-of-the-art GPT-3.5 model to generate such information along with domain-general conversations and facial expressions based on prompt engineering.

Alessandro Suglia, Yonatan Bisk, Ioannis Konstas et al.

Guessing games are a prototypical instance of the "learning by interacting" paradigm. This work investigates how well an artificial agent can benefit from playing guessing games when later asked to perform on novel NLP downstream tasks such as Visual Question Answering (VQA). We propose two ways to exploit playing guessing games: 1) a supervised learning scenario in which the agent learns to mimic successful guessing games and 2) a novel way for an agent to play by itself, called Self-play via Iterated Experience Learning (SPIEL). We evaluate the ability of both procedures to generalize: an in-domain evaluation shows an increased accuracy (+7.79) compared with competitors on the evaluation suite CompGuessWhat?!; a transfer evaluation shows improved performance for VQA on the TDIUC dataset in terms of harmonic average accuracy (+5.31) thanks to more fine-grained object representations learned via SPIEL.

Emanuele Bastianelli, Andrea Vanzo, Oliver Lemon

We study the problem of integrating syntactic information from constituency trees into a neural model in Frame-semantic parsing sub-tasks, namely Target Identification (TI), FrameIdentification (FI), and Semantic Role Labeling (SRL). We use a Graph Convolutional Network to learn specific representations of constituents, such that each constituent is profiled as the production grammar rule it corresponds to. We leverage these representations to build syntactic features for each word in a sentence, computed as the sum of all the constituents on the path between a word and a task-specific node in the tree, e.g. the target predicate for SRL. Our approach improves state-of-the-art results on the TI and SRL of ~1%and~3.5% points, respectively (+2.5% additional points are gained with BERT as input), when tested on FrameNet 1.5, while yielding comparable results on the CoNLL05 dataset to other syntax-aware systems.

Alessandro Suglia, Antonio Vergari, Ioannis Konstas et al.

In visual guessing games, a Guesser has to identify a target object in a scene by asking questions to an Oracle. An effective strategy for the players is to learn conceptual representations of objects that are both discriminative and expressive enough to ask questions and guess correctly. However, as shown by Suglia et al. (2020), existing models fail to learn truly multi-modal representations, relying instead on gold category labels for objects in the scene both at training and inference time. This provides an unnatural performance advantage when categories at inference time match those at training time, and it causes models to fail in more realistic "zero-shot" scenarios where out-of-domain object categories are involved. To overcome this issue, we introduce a novel "imagination" module based on Regularized Auto-Encoders, that learns context-aware and category-aware latent embeddings without relying on category labels at inference time. Our imagination module outperforms state-of-the-art competitors by 8.26% gameplay accuracy in the CompGuessWhat?! zero-shot scenario (Suglia et al., 2020), and it improves the Oracle and Guesser accuracy by 2.08% and 12.86% in the GuessWhat?! benchmark, when no gold categories are available at inference time. The imagination module also boosts reasoning about object properties and attributes.

Daniel Hernández García, Yanchao Yu, Weronika Sieińska et al.

In this paper, we propose a minimum set of concepts and signals needed to track the social state during Human-Robot Interaction. We look into the problem of complex continuous interactions in a social context with multiple humans and robots, and discuss the creation of an explainable and tractable representation/model of their social interaction. We discuss these representations according to their representational and communicational properties, and organize them into four cognitive domains (scene-understanding, behaviour-profiling, mental-state, and dialogue-grounding).

Alessandro Suglia, Ioannis Konstas, Andrea Vanzo et al.

Approaches to Grounded Language Learning typically focus on a single task-based final performance measure that may not depend on desirable properties of the learned hidden representations, such as their ability to predict salient attributes or to generalise to unseen situations. To remedy this, we present GROLLA, an evaluation framework for Grounded Language Learning with Attributes with three sub-tasks: 1) Goal-oriented evaluation; 2) Object attribute prediction evaluation; and 3) Zero-shot evaluation. We also propose a new dataset CompGuessWhat?! as an instance of this framework for evaluating the quality of learned neural representations, in particular concerning attribute grounding. To this end, we extend the original GuessWhat?! dataset by including a semantic layer on top of the perceptual one. Specifically, we enrich the VisualGenome scene graphs associated with the GuessWhat?! images with abstract and situated attributes. By using diagnostic classifiers, we show that current models learn representations that are not expressive enough to encode object attributes (average F1 of 44.27). In addition, they do not learn strategies nor representations that are robust enough to perform well when novel scenes or objects are involved in gameplay (zero-shot best accuracy 50.06%).

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.

Andrea Vanzo, Emanuele Bastianelli, Oliver Lemon

We present a new neural architecture for wide-coverage Natural Language Understanding in Spoken Dialogue Systems. We develop a hierarchical multi-task architecture, which delivers a multi-layer representation of sentence meaning (i.e., Dialogue Acts and Frame-like structures). The architecture is a hierarchy of self-attention mechanisms and BiLSTM encoders followed by CRF tagging layers. We describe a variety of experiments, showing that our approach obtains promising results on a dataset annotated with Dialogue Acts and Frame Semantics. Moreover, we demonstrate its applicability to a different, publicly available NLU dataset annotated with domain-specific intents and corresponding semantic roles, providing overall performance higher than state-of-the-art tools such as RASA, Dialogflow, LUIS, and Watson. For example, we show an average 4.45% improvement in entity tagging F-score over Rasa, Dialogflow and LUIS.

Mary Ellen Foster, Bart Craenen, Amol Deshmukh et al.

In the EU-funded MuMMER project, we have developed a social robot designed to interact naturally and flexibly with users in public spaces such as a shopping mall. We present the latest version of the robot system developed during the project. This system encompasses audio-visual sensing, social signal processing, conversational interaction, perspective taking, geometric reasoning, and motion planning. It successfully combines all these components in an overarching framework using the Robot Operating System (ROS) and has been deployed to a shopping mall in Finland interacting with customers. In this paper, we describe the system components, their interplay, and the resulting robot behaviours and scenarios provided at the shopping mall.

Christian Dondrup, Ioannis Papaioannou, Oliver Lemon

Smart speakers and robots become ever more prevalent in our daily lives. These agents are able to execute a wide range of tasks and actions and, therefore, need systems to control their execution. Current state-of-the-art such as (deep) reinforcement learning, however, requires vast amounts of data for training which is often hard to come by when interacting with humans. To overcome this issue, most systems still rely on Finite State Machines. We introduce Petri Net Machines which present a formal definition for state machines based on Petri Nets that are able to execute concurrent actions reliably, execute and interleave several plans at the same time, and provide an easy to use modelling language. We show their workings based on the example of Human-Robot Interaction in a shopping mall.

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.

Igor Shalyminov, Ondřej Dušek, Oliver Lemon

The overall objective of 'social' dialogue systems is to support engaging, entertaining, and lengthy conversations on a wide variety of topics, including social chit-chat. Apart from raw dialogue data, user-provided ratings are the most common signal used to train such systems to produce engaging responses. In this paper we show that social dialogue systems can be trained effectively from raw unannotated data. Using a dataset of real conversations collected in the 2017 Alexa Prize challenge, we developed a neural ranker for selecting 'good' system responses to user utterances, i.e. responses which are likely to lead to long and engaging conversations. We show that (1) our neural ranker consistently outperforms several strong baselines when trained to optimise for user ratings; (2) when trained on larger amounts of data and only using conversation length as the objective, the ranker performs better than the one trained using ratings -- ultimately reaching a Precision@1 of 0.87. This advance will make data collection for social conversational agents simpler and less expensive in the future.

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.

Ioannis Papaioannou, Amanda Cercas Curry, Jose L. Part et al.

Open-domain social dialogue is one of the long-standing goals of Artificial Intelligence. This year, the Amazon Alexa Prize challenge was announced for the first time, where real customers get to rate systems developed by leading universities worldwide. The aim of the challenge is to converse "coherently and engagingly with humans on popular topics for 20 minutes". We describe our Alexa Prize system (called 'Alana') consisting of an ensemble of bots, combining rule-based and machine learning systems, and using a contextual ranking mechanism to choose a system response. The ranker was trained on real user feedback received during the competition, where we address the problem of how to train on the noisy and sparse feedback obtained during the competition.

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.

Jekaterina Novikova, Christian Dondrup, Ioannis Papaioannou et al.

Recognition of social signals, from human facial expressions or prosody of speech, is a popular research topic in human-robot interaction studies. There is also a long line of research in the spoken dialogue community that investigates user satisfaction in relation to dialogue characteristics. However, very little research relates a combination of multimodal social signals and language features detected during spoken face-to-face human-robot interaction to the resulting user perception of a robot. In this paper we show how different emotional facial expressions of human users, in combination with prosodic characteristics of human speech and features of human-robot dialogue, correlate with users' impressions of the robot after a conversation. We find that happiness in the user's recognised facial expression strongly correlates with likeability of a robot, while dialogue-related features (such as number of human turns or number of sentences per robot utterance) correlate with perceiving a robot as intelligent. In addition, we show that facial expression, emotional features, and prosody are better predictors of human ratings related to perceived robot likeability and anthropomorphism, while linguistic and non-linguistic features more often predict perceived robot intelligence and interpretability. As such, these characteristics may in future be used as an online reward signal for in-situ Reinforcement Learning based adaptive human-robot dialogue systems.

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.

Jekaterina Novikova, Oliver Lemon, Verena Rieser

Recent advances in corpus-based Natural Language Generation (NLG) hold the promise of being easily portable across domains, but require costly training data, consisting of meaning representations (MRs) paired with Natural Language (NL) utterances. In this work, we propose a novel framework for crowdsourcing high quality NLG training data, using automatic quality control measures and evaluating different MRs with which to elicit data. We show that pictorial MRs result in better NL data being collected than logic-based MRs: utterances elicited by pictorial MRs are judged as significantly more natural, more informative, and better phrased, with a significant increase in average quality ratings (around 0.5 points on a 6-point scale), compared to using the logical MRs. As the MR becomes more complex, the benefits of pictorial stimuli increase. The collected data will be released as part of this submission.

Verena Rieser, Oliver Lemon

We present and evaluate a new model for Natural Language Generation (NLG) in Spoken Dialogue Systems, based on statistical planning, given noisy feedback from the current generation context (e.g. a user and a surface realiser). We study its use in a standard NLG problem: how to present information (in this case a set of search results) to users, given the complex trade- offs between utterance length, amount of information conveyed, and cognitive load. We set these trade-offs by analysing existing MATCH data. We then train a NLG pol- icy using Reinforcement Learning (RL), which adapts its behaviour to noisy feed- back from the current generation context. This policy is compared to several base- lines derived from previous work in this area. The learned policy significantly out- performs all the prior approaches.

Dimitra Gkatzia, Oliver Lemon, Verena Rieser

Decision-making is often dependent on uncertain data, e.g. data associated with confidence scores or probabilities. We present a comparison of different information presentations for uncertain data and, for the first time, measure their effects on human decision-making. We show that the use of Natural Language Generation (NLG) improves decision-making under uncertainty, compared to state-of-the-art graphical-based representation methods. In a task-based study with 442 adults, we found that presentations using NLG lead to 24% better decision-making on average than the graphical presentations, and to 44% better decision-making when NLG is combined with graphics. We also show that women achieve significantly better results when presented with NLG output (an 87% increase on average compared to graphical presentations).

Heriberto Cuayáhuitl, Simon Keizer, Oliver Lemon

Artificially intelligent agents equipped with strategic skills that can negotiate during their interactions with other natural or artificial agents are still underdeveloped. This paper describes a successful application of Deep Reinforcement Learning (DRL) for training intelligent agents with strategic conversational skills, in a situated dialogue setting. Previous studies have modelled the behaviour of strategic agents using supervised learning and traditional reinforcement learning techniques, the latter using tabular representations or learning with linear function approximation. In this study, we apply DRL with a high-dimensional state space to the strategic board game of Settlers of Catan---where players can offer resources in exchange for others and they can also reply to offers made by other players. Our experimental results report that the DRL-based learnt policies significantly outperformed several baselines including random, rule-based, and supervised-based behaviours. The DRL-based policy has a 53% win rate versus 3 automated players (`bots'), whereas a supervised player trained on a dialogue corpus in this setting achieved only 27%, versus the same 3 bots. This result supports the claim that DRL is a promising framework for training dialogue systems, and strategic agents with negotiation abilities.

Zhuoran Wang, Paul A. Crook, Wenshuo Tang et al.

Belief compression improves the tractability of large-scale partially observable Markov decision processes (POMDPs) by finding projections from high-dimensional belief space onto low-dimensional approximations, where solving to obtain action selection policies requires fewer computations. This paper develops a unified theoretical framework to analyse three existing linear belief compression approaches, including value-directed compression and two non-negative matrix factorisation (NMF) based algorithms. The results indicate that all the three known belief compression methods have their own critical deficiencies. Therefore, projective NMF belief compression is proposed (P-NMF), aiming to overcome the drawbacks of the existing techniques. The performance of the proposed algorithm is examined on four POMDP problems of reasonably large scale, in comparison with existing techniques. Additionally, the competitiveness of belief compression is compared empirically to a state-of-the-art heuristic search based POMDP solver and their relative merits in solving large-scale POMDPs are investigated.