ReproHum is a large multi-institution project designed to examine the reproducibility of human evaluations of natural language processing. As part of the second phase of the project, we attempt to reproduce an evaluation of the fluency of continuations generated by a pre-trained language model compared to a range of baselines. Working within the constraints of the project, with limited information about the original study, and without access to their participant pool, or the responses of individual participants, we find that we are not able to reproduce the original results. Our participants display a greater tendency to prefer one of the system responses, avoiding a judgement of ‘equal fluency’ more than in the original study. We also conduct further evaluations: we elicit ratings from (1) a broader range of participants; (2) from the same participants at different times; and (3) with an altered definition of fluency. Results of these experiments suggest that the original evaluation collected too few ratings, and that the task formulation may be quite ambiguous. Overall, although we were able to conduct a re-evaluation study, we conclude that the original evaluation was not comprehensive enough to make truly meaningful comparisons

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.

Adversarial robustness evaluates the worst-case performance scenario of a machine learning model to ensure its safety and reliability. For example, cases where the user input contains a minimal change, e.g. a synonym, which causes the previously correct model to return a wrong answer. Using this scenario, this study is the first to investigate the robustness of visually grounded dialog models towards textual attacks. We first aim to understand how multimodal input components contribute to model robustness. Our results show that models which encode dialog history are more robust by providing redundant information. This is in contrast to prior work which finds that dialog history is negligible for model performance on this task. We also evaluate how to generate adversarial test examples which successfully fool the model but remain undetected by the user/software designer. Our analysis shows that the textual, as well as the visual context are important to generate plausible attacks.

Online Gender-Based Violence (GBV), such as misogynistic abuse is an increasingly prevalent problem that technological approaches have struggled to address. Through the lens of the GBV framework, which is rooted in social science and policy, we systematically review 63 available resources for automated identification of such language. We find the datasets are limited in a number of important ways, such as their lack of theoretical grounding and stakeholder input, static nature, and focus on certain media platforms. Based on this review, we recommend development of future resources rooted in sociological expertise andcentering stakeholder voices, namely GBV experts and people with lived experience of GBV.

There are two competing approaches for modelling annotator disagreement: distributional soft-labelling approaches (which aim to capture the level of disagreement) or modelling perspectives of individual annotators or groups thereof. We adapt a multi-task architecture which has previously shown success in modelling perspectives to evaluate its performance on the SEMEVAL Task 11. We do so by combining both approaches, i.e. predicting individual annotator perspectives as an interim step towards predicting annotator disagreement. Despite its previous success, we found that a multi-task approach performed poorly on datasets which contained distinct annotator opinions, suggesting that this approach may not always be suitable when modelling perspectives. Furthermore, our results explain that while strongly perspectivist approaches might not achieve state-of-the-art performance according to evaluation metrics used by distributional approaches, our approach allows for a more nuanced understanding of individual perspectives present in the data. We argue that perspectivist approaches are preferable because they enable decision makers to amplify minority views, and that it is important to re-evaluate metrics to reflect this goal.

NLP datasets annotated with human judgments are rife with disagreements between the judges. This is especially true for tasks depending on subjective judgments such as sentiment analysis or offensive language detection. Particularly in these latter cases, the NLP community has come to realize that the common approach of reconciling’ these different subjective interpretations risks misrepresenting the evidence. Many NLP researchers have therefore concluded that rather than eliminating disagreements from annotated corpora, we should preserve themindeed, some argue that corpora should aim to preserve all interpretations produced by annotators. But this approach to corpus creation for NLP has not yet been widely accepted. The objective of the Le-Wi-Di series of shared tasks is to promote this approach to developing NLP models by providing a unified framework for training and evaluating with such datasets. We report on the second such shared task, which differs from the first edition in three crucial respects: (i) it focuses entirely on NLP, instead of both NLP and computer vision tasks in its first edition; (ii) it focuses on subjective tasks, instead of covering different types of disagreements as training with aggregated labels for subjective NLP tasks is in effect a misrepresentation of the data; and (iii) for the evaluation, we concentrated on soft approaches to evaluation. This second edition of Le-Wi-Di attracted a wide array of partici- pants resulting in 13 shared task submission papers.

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.

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.

Automated dialogue or conversational systems are anthropomorphised by developers and personified by users. While a degree of anthropomorphism is inevitable, conscious and unconscious design choices can guide users to personify them to varying degrees. Encouraging users to relate to automated systems as if they were human can lead to transparency and trust issues, and high risk scenarios caused by over-reliance on their outputs. As a result, natural language processing researchers have investigated the factors that induce personification and develop resources to mitigate such effects. However, these efforts are fragmented, and many aspects of anthropomorphism have yet to be explored. In this paper, we discuss the linguistic factors that contribute to the anthropomorphism of dialogue systems and the harms that can arise thereof, including reinforcing gender stereotypes and conceptions of acceptable language. We recommend that future efforts towards developing dialogue systems take particular care in their design, development, release, and description; and attend to the many linguistic cues that can elicit personification by users.

We report our efforts in identifying a set of previous human evaluations in NLP that would be suitable for a coordinated study examining what makes human evaluations in NLP more/less reproducible. We present our results and findings, which include that just 13% of papers had (i) sufficiently low barriers to reproduction, and (ii) enough obtainable information, to be considered for reproduction, and that all but one of the experiments we selected for reproduction was discovered to have flaws that made the meaningfulness of conducting a reproduction questionable. As a result, we had to change our coordinated study design from a reproduce approach to a standardise-then-reproduce-twice approach. Our overall (negative) finding that the great majority of human evaluations in NLP is not repeatable and/or not reproducible and/or too flawed to justify reproduction, paints a dire picture, but presents an opportunity for a rethink about how to design and report human evaluations in NLP.

Conversational AI systems can engage in unsafe behaviour when handling users’ medical queries that may have severe consequences and could even lead to deaths. Systems therefore need to be capable of both recognising the seriousness of medical inputs and producing responses with appropriate levels of risk. We create a corpus of human written English language medical queries and the responses of different types of systems. We label these with both crowdsourced and expert annotations. While individual crowdworkers may be unreliable at grading the seriousness of the prompts, their aggregated labels tend to agree with professional opinion to a greater extent on identifying the medical queries and recognising the risk types posed by the responses. Results of classification experiments suggest that, while these tasks can be automated, caution should be exercised, as errors can potentially be very serious.

Over the last several years, end-to-end neural conversational agents have vastly improved their ability to carry unrestricted, open-domain conversations with humans. However, these models are often trained on large datasets from the Internet and, as a result, may learn undesirable behaviours from this data, such as toxic or otherwise harmful language. Thus, researchers must wrestle with how and when to release these models. In this paper, we survey recent and related work to highlight tensions between values, potential positive impact, and potential harms. We also provide a framework to support practitioners in deciding whether and how to release these models, following the tenets of value-sensitive design.

We demonstrate EMMA, an embodied multimodal agent which has been developed for the Alexa Prize SimBot challenge. The agent acts within a 3D simulated environment for household tasks. EMMA is a unified and multimodal generative model aimed at solving embodied tasks. In contrast to previous work, our approach treats multiple multimodal tasks as a single multimodal conditional text generation problem, where a model learns to output text given both language and visual input. Furthermore, we showcase that a single generative agent can solve tasks with visual inputs of varying length, such as answering questions about static images, or executing actions given a sequence of previous frames and dialogue utterances. The demo system will allow users to interact conversationally with EMMA in embodied dialogues in different 3D environments from the TEACh dataset.

The social impact of natural language processing and its applications has received increasing attention. In this position paper, we focus on the problem of safety for end-to-end conversational AI. We survey the problem landscape therein, introducing a taxonomy of three observed phenomena: the Instigator, Yea-Sayer, and Impostor effects. We then empirically assess the extent to which current tools can measure these effects and current systems display them. We release these tools as part of a “first aid kit” (SafetyKit) to quickly assess apparent safety concerns. Our results show that, while current tools are able to provide an estimate of the relative safety of systems in various settings, they still have several shortcomings. We suggest several future directions and discuss ethical considerations.

Technology companies have produced varied responses to concerns about the effects of the design of their conversational AI systems. Some have claimed that their voice assistants are in fact not gendered or human-like—despite design features suggesting the contrary. We compare these claims to user perceptions by analysing the pronouns they use when referring to AI assistants. We also examine systems’ responses and the extent to which they generate output which is gendered and anthropomorphic. We find that, while some companies appear to be addressing the ethical concerns raised, in some cases, their claims do not seem to hold true. In particular, our results show that system outputs are ambiguous as to the humanness of the systems, and that users tend to personify and gender them as a result.

We present the first English corpus study on abusive language towards three conversational AI systems gathered ‘in the wild’: an open-domain social bot, a rule-based chatbot, and a task-based system. To account for the complexity of the task, we take a more ‘nuanced’ approach where our ConvAI dataset reflects fine-grained notions of abuse, as well as views from multiple expert annotators. We find that the distribution of abuse is vastly different compared to other commonly used datasets, with more sexually tinted aggression towards the virtual persona of these systems. Finally, we report results from bench-marking existing models against this data. Unsurprisingly, we find that there is substantial room for improvement with F1 scores below 90%.

Previous work has shown that human evaluations in NLP are notoriously under-powered. Here, we argue that there are two common factors which make this problem even worse: NLP studies usually (a) treat ordinal data as interval data and (b) operate under high variance settings while the differences they are hoping to detect are often subtle. We demonstrate through simulation that ordinal mixed effects models are better able to detect small differences between models, especially in high variance settings common in evaluations of generated texts. We release tools for researchers to conduct their own power analysis and test their assumptions. We also make recommendations for improving statistical power.

One of the most challenging aspects of current single-document news summarization is that the summary often contains ‘extrinsic hallucinations’, i.e., facts that are not present in the source document, which are often derived via world knowledge. This causes summarisation systems to act more like open-ended language models tending to hallucinate facts that are erroneous. In this paper, we mitigate this problem with the help of multiple supplementary resource documents assisting the task. We present a new dataset MiraNews and benchmark existing summarisation models. In contrast to multi-document summarization, which addresses multiple events from several source documents, we still aim at generating a summary for a single document. We show via data analysis that it’s not only the models which are to blame: more than 27% of facts mentioned in the gold summaries of MiraNews are better grounded on assisting documents than in the main source articles. An error analysis of generated summaries from pretrained models fine-tuned on MIRANEWS reveals that this has an even bigger effects on models: assisted summarisation reduces 55% of hallucinations when compared to single-document summarisation models trained on the main article only.

We present AggGen (pronounced ‘again’) a data-to-text model which re-introduces two explicit sentence planning stages into neural data-to-text systems: input ordering and input aggregation. In contrast to previous work using sentence planning, our model is still end-to-end: AggGen performs sentence planning at the same time as generating text by learning latent alignments (via semantic facts) between input representation and target text. Experiments on the WebNLG and E2E challenge data show that by using fact-based alignments our approach is more interpretable, expressive, robust to noise, and easier to control, while retaining the advantages of end-to-end systems in terms of fluency. Our code is available at https://github.com/XinnuoXu/AggGen.

Mixed initiative in open-domain dialogue requires a system to pro-actively introduce new topics. The one-turn topic transition task explores how a system connects two topics in a cooperative and coherent manner. The goal of the task is to generate a “bridging” utterance connecting the new topic to the topic of the previous conversation turn. We are especially interested in commonsense explanations of how a new topic relates to what has been mentioned before. We first collect a new dataset of human one-turn topic transitions, which we callOTTers. We then explore different strategies used by humans when asked to complete such a task, and notice that the use of a bridging utterance to connect the two topics is the approach used the most. We finally show how existing state-of-the-art text generation models can be adapted to this task and examine the performance of these baselines on different splits of the OTTers data.

Human assessment remains the most trusted form of evaluation in NLG, but highly diverse approaches and a proliferation of different quality criteria used by researchers make it difficult to compare results and draw conclusions across papers, with adverse implications for meta-evaluation and reproducibility. In this paper, we present (i) our dataset of 165 NLG papers with human evaluations, (ii) the annotation scheme we developed to label the papers for different aspects of evaluations, (iii) quantitative analyses of the annotations, and (iv) a set of recommendations for improving standards in evaluation reporting. We use the annotations as a basis for examining information included in evaluation reports, and levels of consistency in approaches, experimental design and terminology, focusing in particular on the 200+ different terms that have been used for evaluated aspects of quality. We conclude that due to a pervasive lack of clarity in reports and extreme diversity in approaches, human evaluation in NLG presents as extremely confused in 2020, and that the field is in urgent need of standard methods and terminology.

Conversational voice assistants are rapidly developing from purely transactional systems to social companions with “personality”. UNESCO recently stated that the female and submissive personality of current digital assistants gives rise for concern as it reinforces gender stereotypes. In this work, we present results from a participatory design workshop, where we invite people to submit their preferences for a what their ideal persona might look like, both in drawings as well as in a multiple choice questionnaire. We find no clear consensus which suggests that one possible solution is to let people configure/personalise their assistants. We then outline a multi-disciplinary project of how we plan to address the complex question of gender and stereotyping in digital assistants.

Abstractive summarisation is notoriously hard to evaluate since standard word-overlap-based metrics are insufficient. We introduce a new evaluation metric which is based on fact-level content weighting, i.e. relating the facts of the document to the facts of the summary. We fol- low the assumption that a good summary will reflect all relevant facts, i.e. the ones present in the ground truth (human-generated refer- ence summary). We confirm this hypothe- sis by showing that our weightings are highly correlated to human perception and compare favourably to the recent manual highlight- based metric of Hardy et al. (2019).

Visual Dialogue involves “understanding” the dialogue history (what has been discussed previously) and the current question (what is asked), in addition to grounding information in the image, to accurately generate the correct response. In this paper, we show that co-attention models which explicitly encode dialoh history outperform models that don’t, achieving state-of-the-art performance (72 % NDCG on val set). However, we also expose shortcomings of the crowdsourcing dataset collection procedure, by showing that dialogue history is indeed only required for a small amount of the data, and that the current evaluation metric encourages generic replies. To that end, we propose a challenging subset (VisdialConv) of the VisdialVal set and the benchmark NDCG of 63%.

Spoken Language Understanding infers semantic meaning directly from audio data, and thus promises to reduce error propagation and misunderstandings in end-user applications. However, publicly available SLU resources are limited. In this paper, we release SLURP, a new SLU package containing the following: (1) A new challenging dataset in English spanning 18 domains, which is substantially bigger and linguistically more diverse than existing datasets; (2) Competitive baselines based on state-of-the-art NLU and ASR systems; (3) A new transparent metric for entity labelling which enables a detailed error analysis for identifying potential areas of improvement. SLURP is available at https://github.com/pswietojanski/slurp.

How should conversational agents respond to verbal abuse through the user? To answer this question, we conduct a large-scale crowd-sourced evaluation of abuse response strategies employed by current state-of-the-art systems. Our results show that some strategies, such as “polite refusal”, score highly across the board, while for other strategies demographic factors, such as age, as well as the severity of the preceding abuse influence the user’s perception of which response is appropriate. In addition, we find that most data-driven models lag behind rule-based or commercial systems in terms of their perceived appropriateness.

We present the first complete spoken dialogue system driven by a multiimensional statistical dialogue manager. This framework has been shown to substantially reduce data needs by leveraging domain-independent dimensions, such as social obligations or feedback, which (as we show) can be transferred between domains. In this paper, we conduct a user study and show that the performance of a multi-dimensional system, which can be adapted from a source domain, is equivalent to that of a one-dimensional baseline, which can only be trained from scratch.

We present a recurrent neural network based system for automatic quality estimation of natural language generation (NLG) outputs, which jointly learns to assign numerical ratings to individual outputs and to provide pairwise rankings of two different outputs. The latter is trained using pairwise hinge loss over scores from two copies of the rating network. We use learning to rank and synthetic data to improve the quality of ratings assigned by our system: We synthesise training pairs of distorted system outputs and train the system to rank the less distorted one higher. This leads to a 12% increase in correlation with human ratings over the previous benchmark. We also establish the state of the art on the dataset of relative rankings from the E2E NLG Challenge (Dusek et al., 2019), where synthetic data lead to a 4% accuracy increase over the base model.

Neural natural language generation (NNLG) systems are known for their pathological outputs, i.e. generating text which is unrelated to the input specification. In this paper, we show the impact of semantic noise on state-of-the-art NNLG models which implement different semantic control mechanisms. We find that cleaned data can improve semantic correctness by up to 97%, while maintaining fluency. We also find that the most common error is omitting information, rather than hallucination.

Conversational AI systems, such as Amazon’s Alexa, are rapidly developing from purely transactional systems to social chatbots, which can respond to a wide variety of user requests. In this article, we establish how current state-of-the-art conversational systems react to inappropriate requests, such as bullying and sexual harassment on the part of the user, by collecting and analysing the novel #MeTooAlexa corpus. Our results show that commercial systems mainly avoid answering, while rule-based chatbots show a variety of behaviours and often deflect. Data-driven systems, on the other hand, are often non-coherent, but also run the risk of being interpreted as flirtatious and sometimes react with counter-aggression. This includes our own system, trained on “clean” data, which suggests that inappropriate system behaviour is not caused by data bias.

Multimodal search-based dialogue is a challenging new task: It extends visually grounded question answering systems into multi-turn conversations with access to an external database. We address this new challenge by learning a neural response generation system from the recently released Multimodal Dialogue (MMD) dataset (Saha et al., 2017). We introduce a knowledge-grounded multimodal conversational model where an encoded knowledge base (KB) representation is appended to the decoder input. Our model substantially outperforms strong baselines in terms of text-based similarity measures (over 9 BLEU points, 3 of which are solely due to the use of additional information from the KB).

In this work, we investigate the task of textual response generation in a multimodal task-oriented dialogue system. Our work is based on the recently released Multimodal Dialogue (MMD) dataset (Saha et al., 2017) in the fashion domain. We introduce a multimodal extension to the Hierarchical Recurrent Encoder-Decoder (HRED) model and show that this extension outperforms strong baselines in terms of text-based similarity metrics. We also showcase the shortcomings of current vision and language models by performing an error analysis on our system’s output.

This paper summarises the experimental setup and results of the first shared task on end-to-end (E2E) natural language generation (NLG) in spoken dialogue systems. Recent end-to-end generation systems are promising since they reduce the need for data annotation. However, they are currently limited to small, delexicalised datasets. The E2E NLG shared task aims to assess whether these novel approaches can generate better-quality output by learning from a dataset containing higher lexical richness, syntactic complexity and diverse discourse phenomena. We compare 62 systems submitted by 17 institutions, covering a wide range of approaches, including machine learning architectures – with the majority implementing sequence-to-sequence models (seq2seq) – as well as systems based on grammatical rules and templates.

Human evaluation for natural language generation (NLG) often suffers from inconsistent user ratings. While previous research tends to attribute this problem to individual user preferences, we show that the quality of human judgements can also be improved by experimental design. We present a novel rank-based magnitude estimation method (RankME), which combines the use of continuous scales and relative assessments. We show that RankME significantly improves the reliability and consistency of human ratings compared to traditional evaluation methods. In addition, we show that it is possible to evaluate NLG systems according to multiple, distinct criteria, which is important for error analysis. Finally, we demonstrate that RankME, in combination with Bayesian estimation of system quality, is a cost-effective alternative for ranking multiple NLG systems.

We present three enhancements to existing encoder-decoder models for open-domain conversational agents, aimed at effectively modeling coherence and promoting output diversity: (1) We introduce a measure of coherence as the GloVe embedding similarity between the dialogue context and the generated response, (2) we filter our training corpora based on the measure of coherence to obtain topically coherent and lexically diverse context-response pairs, (3) we then train a response generator using a conditional variational autoencoder model that incorporates the measure of coherence as a latent variable and uses a context gate to guarantee topical consistency with the context and promote lexical diversity. Experiments on the OpenSubtitles corpus show a substantial improvement over competitive neural models in terms of BLEU score as well as metrics of coherence and diversity.

This paper describes the E2E data, a new dataset for training end-to-end, data-driven natural language generation systems in the restaurant domain, which is ten times bigger than existing, frequently used datasets in this area. The E2E dataset poses new challenges: (1) its human reference texts show more lexical richness and syntactic variation, including discourse phenomena; (2) generating from this set requires content selection. As such, learning from this dataset promises more natural, varied and less template-like system utterances. We also establish a baseline on this dataset, which illustrates some of the difficulties associated with this data.

The majority of NLG evaluation relies on automatic metrics, such as BLEU . In this paper, we motivate the need for novel, system- and data-independent automatic evaluation methods: We investigate a wide range of metrics, including state-of-the-art word-based and novel grammar-based ones, and demonstrate that they only weakly reflect human judgements of system outputs as generated by data-driven, end-to-end NLG. We also show that metric performance is data- and system-specific. Nevertheless, our results also suggest that automatic metrics perform reliably at system-level and can support system development by finding cases where a system performs poorly.

Our interest is in people’s capacity to efficiently and effectively describe geographic objects in urban scenes. The broader ambition is to develop spatial models capable of equivalent functionality able to construct such referring expressions. To that end we present a newly crowd-sourced data set of natural language references to objects anchored in complex urban scenes (In short: The REAL Corpus ― Referring Expressions Anchored Language). The REAL corpus contains a collection of images of real-world urban scenes together with verbal descriptions of target objects generated by humans, paired with data on how successful other people were able to identify the same object based on these descriptions. In total, the corpus contains 32 images with on average 27 descriptions per image and 3 verifications for each description. In addition, the corpus is annotated with a variety of linguistically motivated features. The paper highlights issues posed by collecting data using crowd-sourcing with an unrestricted input format, as well as using real-world urban scenes.

We present a newly collected data set of 8,868 gold-standard annotated Arabic feeds. The corpus is manually labelled for subjectivity and sentiment analysis (SSA) ( = 0:816). In addition, the corpus is annotated with a variety of motivated feature-sets that have previously shown positive impact on performance. The paper highlights issues posed by twitter as a genre, such as mixture of language varieties and topic-shifts. Our next step is to extend the current corpus, using online semi-supervised learning. A first sub-corpus will be released via the ELRA repository as part of this submission.

The ultimate goal when building dialogue systems is to satisfy the needs of real users, but quality assurance for dialogue strategies is a non-trivial problem. The applied evaluation metrics and resulting design principles are often obscure, emerge by trial-and-error, and are highly context dependent. This paper introduces data-driven methods for obtaining reliable objective functions for system design. In particular, we test whether an objective function obtained from Wizard-of-Oz (WOZ) data is a valid estimate of real users’ preferences. We test this in a test-retest comparison between the model obtained from the WOZ study and the models obtained when testing with real users. We can show that, despite a low fit to the initial data, the objective function obtained from WOZ data makes accurate predictions for automatic dialogue evaluation, and, when automatically optimising a policy using these predictions, the improvement over a strategy simply mimicking the data becomes clear from an error analysis.

We describe a corpus of multimodal dialogues with an MP3player collected in Wizard-of-Oz experiments and annotated with a richfeature set at several layers. We are using the Nite XML Toolkit (NXT) to represent and further process the data. We designed an NXTdata model, converted experiment log file data and manualtranscriptions into NXT, and are building tools for additionalannotation using NXT libraries. The annotated corpus will be used to (i) investigate various aspects of multimodal presentation andinteraction strategies both within and across annotation layers; (ii) design an initial policy for reinforcement learning of multimodalclarification requests.