Hate speech detection is a prevalent research field, yet it remains underexplored at the level of word meaning. This is significant, as terms used to convey hate often involve non-standard or novel usages which might be overlooked by commonly leveraged LMs trained on general language use. In this paper, we introduce the Hateful Word in Context Classification (HateWiC) task and present a dataset of ~4000 WiC-instances, each labeled by three annotators. Our analyses and computational exploration focus on the interplay between the subjective nature (context-dependent connotations) and the descriptive nature (as described in dictionary definitions) of hateful word senses. HateWiC annotations confirm that hatefulness of a word in context does not always derive from the sense definition alone. We explore the prediction of both majority and individual annotator labels, and we experiment with modeling context- and sense-based inputs. Our findings indicate that including definitions proves effective overall, yet not in cases where hateful connotations vary. Conversely, including annotator demographics becomes more important for mitigating performance drop in subjective hate prediction.

Hate speech is a complex and subjective phenomenon. In this paper, we present a dataset (GAZE4HATE) that provides gaze data collected in a hate speech annotation experiment. We study whether the gaze of an annotator provides predictors of their subjective hatefulness rating, and how gaze features can improve Hate Speech Detection (HSD). We conduct experiments on statistical modeling of subjective hate ratings and gaze and analyze to what extent rationales derived from hate speech models correspond to human gaze and explanations in our data. Finally, we introduce MEANION, a first gaze-integrated HSD model. Our experiments show that particular gaze features like dwell time or fixation counts systematically correlate with annotators’ subjective hate ratings and improve predictions of text-only hate speech models.

We propose an end-to-end differentiable training paradigm for stable training of a rationalized transformer classifier. Our approach results in a single model that simultaneously classifies a sample and scores input tokens based on their relevance to the classification. To this end, we build on the widely-used three-player-game for training rationalized models, which typically relies on training a rationale selector, a classifier and a complement classifier. We simplify this approach by making a single model fulfill all three roles, leading to a more efficient training paradigm that is not susceptible to the common training instabilities that plague existing approaches. Further, we extend this paradigm to produce class-wise rationales while incorporating recent advances in parameterizing and regularizing the resulting rationales, thus leading to substantially improved and state-of-the-art alignment with human annotations without any explicit supervision.

We examine how users perceive the limitations of an AI system when it encounters a task that it cannot perform perfectly and whether providing explanations alongside its answers aids users in constructing an appropriate mental model of the system’s capabilities and limitations. We employ a visual question answer and explanation task where we control the AI system’s limitations by manipulating the visual inputs: during inference, the system either processes full-color or grayscale images. Our goal is to determine whether participants can perceive the limitations of the system. We hypothesize that explanations will make limited AI capabilities more transparent to users. However, our results show that explanations do not have this effect. Instead of allowing users to more accurately assess the limitations of the AI system, explanations generally increase users’ perceptions of the system’s competence – regardless of its actual performance.

Natural Language Generation (NLG), and more generally generative AI, are among the currently most impactful research fields. Creative NLG, such as automatic poetry generation, is a fascinating niche in this area. While most previous research has focused on forms of the Turing test when evaluating automatic poetry generation — can humans distinguish between automatic and human generated poetry — we evaluate the diversity of automatically generated poetry (with a focus on quatrains), by comparing distributions of generated poetry to distributions of human poetry along structural, lexical, semantic and stylistic dimensions, assessing different model types (word vs. character-level, general purpose LLMs vs. poetry-specific models), including the very recent LLaMA3-8B, and types of fine-tuning (conditioned vs. unconditioned). We find that current automatic poetry systems are considerably underdiverse along multiple dimensions — they often do not rhyme sufficiently, are semantically too uniform and even do not match the length distribution of human poetry. Our experiments reveal, however, that style-conditioning and character-level modeling clearly increases diversity across virtually all dimensions we explore. Our identified limitations may serve as the basis for more genuinely diverse future poetry generation models.

Syntactic learning curves in LMs are usually reported as relatively stable and power law-shaped. By analyzing the learning curves of different LMs on various syntactic phenomena using both small self-trained llama models and larger pre-trained pythia models, we show that while many phenomena do follow typical power law curves, others exhibit S-shaped, U-shaped, or erratic patterns. Certain syntactic paradigms remain challenging even for large models, resulting in persistent preference for ungrammatical sentences. Most phenomena show similar curves for their paradigms, but the existence of diverging patterns and oscillations indicates that average curves mask important developments, underscoring the need for more detailed analyses of individual learning trajectories.

Scene context is well known to facilitate humans’ perception of visible objects. In this paper, we investigate the role of context in Referring Expression Generation (REG) for objects in images, where existing research has often focused on distractor contexts that exert pressure on the generator. We take a new perspective on scene context in REG and hypothesize that contextual information can be conceived of as a resource that makes REG models more resilient and facilitates the generation of object descriptions, and object types in particular. We train and test Transformer-based REG models with target representations that have been artificially obscured with noise to varying degrees. We evaluate how properties of the models’ visual context affect their processing and performance. Our results show that even simple scene contexts make models surprisingly resilient to perturbations, to the extent that they can identify referent types even when visual information about the target is completely missing.

Research in Language & Vision rarely uses naturally occurring multimodal documents as Wikipedia articles, since they feature complex image-text relations and implicit image-text alignments. In this paper, we provide one of the first datasets that provides ground-truth annotations of image-text alignments in multi-paragraph multi-image articles. The dataset can be used to study phenomena of visual language grounding in longer documents and assess retrieval capabilities of language models trained on, e.g., captioning data. Our analyses show that there are systematic linguistic differences between the image captions and descriptive sentences from the article’s text and that intra-document retrieval is a challenging task for state-of-the-art models in L&V (CLIP, VILT, MCSE).

We introduce SlayQA, a novel benchmark data set designed to evaluate language models’ ability to handle gender-inclusive language, specifically the use of neopronouns, in a question-answering setting. Derived from the Social IQa data set, SlayQA modifies context-question-answer triples to include gender-neutral pronouns, creating a significant linguistic distribution shift in comparison to common pre-training corpora like C4 or Dolma. Our results show that state-of-the-art language models struggle with the challenge, exhibiting small, but noticeable performance drops when answering question containing neopronouns compared to those without.

Using Brennan and Clark’s theory of a Conceptual Pact, that when interlocutors agree on a name for an object, they are forming a temporary agreement on how to conceptualize that object, we present an extension to a simple reference resolver which simulates this process over time with different conversation pairs. In a puzzle construction domain, we model pacts with small language models for each referent which update during the interaction. When features from these pact models are incorporated into a simple bag-of-words reference resolver, the accuracy increases compared to using a standard pre-trained model. The model performs equally to a competitor using the same data but with exhaustive re-training after each prediction, while also being more transparent, faster and less resource-intensive. We also experiment with reducing the number of training interactions, and can still achieve reference resolution accuracies of over 80% in testing from observing a single previous interaction, over 20% higher than a pre-trained baseline. While this is a limited domain, we argue the model could be applicable to larger real-world applications in human and human-robot interaction and is an interpretable and transparent model.

Generating visualizations from natural language queries is a useful extension to visualization libraries such as Vega-Lite. The goal of the NL2VIS task is to generate a valid Vega-Lite specification from a data frame and a natural language query as input, which can then be rendered as a visualization. To enable real-time interaction with the data, small model sizes and fast inferences are required. Previous work has introduced custom neural network solutions with custom visualization specifications and has not systematically tested pre-trained LMs to solve this problem. In this work, we opt for a more generic approach that (i) evaluates pre-trained LMs of different sizes and (ii) uses string encodings of data frames and visualization specifications instead of custom specifications. In our experiments, we show that these representations, in combination with pre-trained LMs, scale better than current state-of-the-art models. In addition, the small and base versions of the T5 architecture achieve real-time interaction, while LLMs far exceed latency thresholds suitable for visual exploration tasks. In summary, our models generate visualization specifications in real-time on a CPU and establish a new state of the art on the NL2VIS benchmark nvBench.

The investigation of lexical change has predominantly focused on generic language evolution, not suited for detecting shifts in a particular domain, such as hate speech. Our study introduces the task of identifying changes in lexical semantics related to hate speech within historical texts. We present an interdisciplinary approach that brings together NLP and History, yielding a pilot dataset comprising 16th-century Early Modern English religious writings during the Protestant Reformation. We provide annotations for both semantic shifts and hatefulness on this data and, thereby, combine the tasks of Lexical Semantic Change Detection and Hate Speech Detection. Our framework and resulting dataset facilitate the evaluation of our applied methods, advancing the analysis of hate speech evolution.

Existing language and vision models achieve impressive performance in image-text understanding. Yet, it is an open question to what extent they can be used for language understanding in 3D environments and whether they implicitly acquire 3D object knowledge, e.g. about different views of an object. In this paper, we investigate whether a state-of-the-art language and vision model, CLIP, is able to ground perspective descriptions of a 3D object and identify canonical views of common objects based on text queries. We present an evaluation framework that uses a circling camera around a 3D object to generate images from different viewpoints and evaluate them in terms of their similarity to natural language descriptions. We find that a pre-trained CLIP model performs poorly on most canonical views and that fine-tuning using hard negative sampling and random contrasting yields good results even under conditions with little available training data.

Concurrent with the rapid progress in neural network-based models in NLP, the need for creating explanations for the predictions of these black-box models has risen steadily. Yet, especially for complex inputs like texts or images, existing interpretability methods still struggle with deriving easily interpretable explanations that also accurately represent the basis for the model’s decision. To this end, we propose a new, model-agnostic method to generate extractive explanations for predictions made by neural networks, that is based on masking parts of the input which the model does not consider to be indicative of the respective class. The masking is done using gradient-based optimization combined with a new regularization scheme that enforces sufficiency, comprehensiveness, and compactness of the generated explanation. Our method achieves state-of-the-art results in a challenging paragraph-level rationale extraction task, showing that this task can be performed without training a specialized model. We further apply our method to image inputs and obtain high-quality explanations for image classifications, which indicates that the objectives for optimizing explanation masks in text generalize to inputs of other modalities.

The prevalence of hate speech on online platforms has become a pressing concern for society, leading to increased attention towards detecting hate speech. Prior work in this area has primarily focused on identifying hate speech at the utterance level that reflects the complex nature of hate speech. In this paper, we propose a targeted and efficient approach to identifying hate speech by detecting slurs at the lexical level using contextualized word embeddings. We hypothesize that slurs have a systematically different representation than their neutral counterparts, making them identifiable through existing methods for discovering semantic dimensions in word embeddings. The results demonstrate the effectiveness of our approach in predicting slurs, confirming linguistic theory that the meaning of slurs is stable across contexts. Our robust hate dimension approach for slur identification offers a promising solution to tackle a smaller yet crucial piece of the complex puzzle of hate speech detection.

Recent studies have used human continuations of Implicit Causality (IC) prompts collected in linguistic experiments to evaluate discourse understanding in large language models (LLMs), focusing on the well-known IC coreference bias in the LLMs’ predictions of the next word following the prompt. In this study, we investigate how continuations of IC prompts can be used to evaluate the text generation capabilities of LLMs in a linguistically controlled setting. We conduct an experiment using two open-source GPT-based models, employing human evaluation to assess different aspects of continuation quality. Our findings show that LLMs struggle in particular with generating coherent continuations in this rather simple setting, indicating a lack of discourse knowledge beyond the well-known IC bias. Our results also suggest that a bias congruent continuation does not necessarily equate to a higher continuation quality. Furthermore, our study draws upon insights from the Uniform Information Density hypothesis, testing different prompt modifications and decoding procedures and showing that sampling-based methods are particularly sensitive to the information density of the prompts.

The success of large language models (LMs) has also prompted a push towards smaller models, but the differences in functionality and encodings between these two types of models are not yet well understood. In this paper, we employ a perturbed masking approach to investigate differences in token influence patterns on the sequence embeddings of larger and smaller RoBERTa models. Specifically, we explore how token properties like position, length or part of speech influence their sequence embeddings. We find that there is a general tendency for sequence-final tokens to exert a higher influence. Among part-of-speech tags, nouns, numerals and punctuation marks are the most influential, with smaller deviations for individual models. These findings also align with usage-based linguistic evidence on the effect of entrenchment. Finally, we show that the relationship between data size and model size influences the variability and brittleness of these effects, hinting towards a need for holistically balanced models.

The increasing interest in probing the linguistic capabilities of large language models (LLMs) has long reached the area of semantics and pragmatics, including the phenomenon of presuppositions. In this study, we investigate a phenomenon that, however, has not yet been investigated, i.e., the phenomenon of anti-presupposition and the principle that accounts for it, the Maximize Presupposition! principle (MP!). Through an experimental investigation using psycholinguistic data and four open-source BERT model variants, we explore how language models handle different anti-presuppositions and whether they apply the MP! principle in their predictions. Further, we examine whether fine-tuning with Natural Language Inference data impacts adherence to the MP! principle. Our findings reveal that LLMs tend to replicate context-based n-grams rather than follow the MP! principle, with fine-tuning not enhancing their adherence. Notably, our results further indicate a striking difficulty of LLMs to correctly predict determiners, in relatively simple linguistic contexts.

In Referring Expression Generation, model inputs are often composed of different representations, including the visual properties of the intended referent, its relative position and size, and the visual context. Yet, the extent to which this information influences the generation process of black-box neural models is largely unclear. We investigate the relative weighting of target, location, and context information in the attention components of a Transformer-based generation model. Our results show a general target bias, which, however, depends on the content of the generated expressions, pointing to interesting directions for future research.

In this research, we investigate whether BERT can differentiate between modal verb senses and sentence modalities and whether it performs equally well on different varieties of English. We fit probing classifiers under two conditions: contextualised embeddings of modal verbs and sentence embeddings. We also investigate BERT’s ability to predict masked modal verbs. Additionally, we classify separately for each modal verb to investigate whether BERT encodes different representations of senses for each individual verb. Lastly, we employ classifiers on data from different varieties of English to determine whether non-American English data is an additional hurdle. Results indicate that BERT has different representations for distinct senses for each modal verb, but does not represent modal sense independently from modal verbs. We also show that performance in different varieties of English is not equal, pointing to a necessary shift in the way we train large language models towards more linguistic diversity. We make our annotated dataset of modal sense in different varieties of English available at https://github.com/wagner-jonas/VEM.

Natural language as a modality of interaction is becoming increasingly popular in the field of visualization. In addition to the popular query interfaces, other language-based interactions such as annotations, recommendations, explanations, or documentation experience growing interest. In this survey, we provide an overview of natural language-based interaction in the research area of visualization. We discuss a renowned taxonomy of visualization tasks and classify 119 related works to illustrate the state-of-the-art of how current natural language interfaces support their performance. We examine applied NLP methods and discuss human-machine dialogue structures with a focus on initiative, duration, and communicative functions in recent visualization-oriented dialogue interfaces. Based on this overview, we point out interesting areas for the future application of NLP methods in the field of visualization.

Although contextualized word embeddings have led to great improvements in automatic language understanding, their potential for practical applications in document exploration and visualization has been little explored. Common visualization techniques used for, e.g., model analysis usually provide simple scatter plots of token-level embeddings that do not provide insight into their contextual use. In this work, we propose KeywordScape, a visual exploration tool that allows to overview, summarize, and explore the semantic content of documents based on their keywords. While existing keyword-based exploration tools assume that keywords have static meanings, our tool represents keywords in terms of their contextualized embeddings. Our application visualizes these embeddings in a semantic landscape that represents keywords as islands on a spherical map. This keeps keywords with similar context close to each other, allowing for a more precise search and comparison of documents.

Referential gaze is a fundamental phenomenon for psycholinguistics and human-human communication. However, modeling referential gaze for real-world scenarios, e.g. for task-oriented communication, is lacking the well-deserved attention from the NLP community. In this paper, we address this challenging issue by proposing a novel multimodal NLP task; namely predicting when the gaze is referential. We further investigate how to model referential gaze and transfer gaze features to adapt to unseen situated settings that target different referential complexities than the training environment. We train (i) a sequential attention-based LSTM model and (ii) a multivariate transformer encoder architecture to predict whether the gaze is on a referent object. The models are evaluated on the three complexity datasets. The results indicate that the gaze features can be transferred not only among various similar tasks and scenes but also across various complexity levels. Taking the referential complexity of a scene into account is important for successful target prediction using gaze parameters especially when there is not much data for fine-tuning.

Referential gaze is a fundamental phenomenon for psycholinguistics and human-human communication. However, modeling referential gaze for real-world scenarios, e.g. for task-oriented communication, is lacking the well-deserved attention from the NLP community. In this paper, we address this challenging issue by proposing a novel multimodal NLP task; namely predicting when the gaze is referential. We further investigate how to model referential gaze and transfer gaze features to adapt to unseen situated settings that target different referential complexities than the training environment. We train (i) a sequential attention-based LSTM model and (ii) a multivariate transformer encoder architecture to predict whether the gaze is on a referent object. The models are evaluated on the three complexity datasets. The results indicate that the gaze features can be transferred not only among various similar tasks and scenes but also across various complexity levels. Taking the referential complexity of a scene into account is important for successful target prediction using gaze parameters especially when there is not much data for fine-tuning.

We investigate the problem of identifying the major hypothesis that is addressed in a scientific paper. To this end, we present a dataset from the domain of invasion biology that organizes a set of 954 papers into a network of fine-grained domain-specific categories of hypotheses. We carry out experiments on classifying abstracts according to these categories and present a pilot study on annotating hypothesis statements within the text. We find that hypothesis statements in our dataset are complex, varied and more or less explicit, and, importantly, spread over the whole abstract. Experiments with BERT-based classifiers show that these models are able to classify complex hypothesis statements to some extent, without being trained on sentence-level text span annotations.

Automatizing the process of understanding the global narrative structure of long texts and stories is still a major challenge for state-of-the-art natural language understanding systems, particularly because annotated data is scarce and existing annotation workflows do not scale well to the annotation of complex narrative phenomena. In this work, we focus on the identification of narrative levels in texts corresponding to stories that are embedded in stories. Lacking sufficient pre-annotated training data, we explore a solution to deal with data scarcity that is common in machine learning: the automatic augmentation of an existing small data set of annotated samples with the help of data synthesis. We present a workflow for narrative level detection, that includes the operationalization of the task, a model, and a data augmentation protocol for automatically generating narrative texts annotated with breaks between narrative levels. Our experiments suggest that narrative levels in long text constitute a challenging phenomenon for state-of-the-art NLP models, but generating training data synthetically does improve the prediction results considerably.

In this work, we explore the fitness of various word/concept representations in analyzing an experimental verbal fluency dataset providing human responses to 10 different category enumeration tasks. Based on human annotations of so-called clusters and switches between sub-categories in the verbal fluency sequences, we analyze whether lexical semantic knowledge represented in word embedding spaces (GloVe, fastText, ConceptNet, BERT) is suitable for detecting these conceptual clusters and switches within and across different categories. Our results indicate that ConceptNet embeddings, a distributional semantics method enriched with taxonomical relations, outperforms other semantic representations by a large margin. Moreover, category-specific analysis suggests that individual thresholds per category are more suited for the analysis of clustering and switching in particular embedding sub-space instead of a one-fits-all cross-category solution. The results point to interesting directions for future work on probing word embedding models on the verbal fluency task.

Multi-modal texts are abundant and diverse in structure, yet Language & Vision research of these naturally occurring texts has mostly focused on genres that are comparatively light on text, like tweets. In this paper, we discuss the challenges and potential benefits of a L&V framework that explicitly models referential relations, taking Wikipedia articles about buildings as an example. We briefly survey existing related tasks in L&V and propose multi-modal information extraction as a general direction for future research.

The shift to neural models in Referring Expression Generation (REG) has enabled more natural set-ups, but at the cost of interpretability. We argue that integrating pragmatic reasoning into the inference of context-agnostic generation models could reconcile traits of traditional and neural REG, as this offers a separation between context-independent, literal information and pragmatic adaptation to context. With this in mind, we apply existing decoding strategies from discriminative image captioning to REG and evaluate them in terms of pragmatic informativity, likelihood to ground-truth annotations and linguistic diversity. Our results show general effectiveness, but a relatively small gain in informativity, raising important questions for REG in general.

Recent work has adopted models of pragmatic reasoning for the generation of informative language in, e.g., image captioning. We propose a simple but highly effective relaxation of fully rational decoding, based on an existing incremental and character-level approach to pragmatically informative neural image captioning. We implement a mixed, ‘fast’ and ‘slow’, speaker that applies pragmatic reasoning occasionally (only word-initially), while unrolling the language model. In our evaluation, we find that increased informativeness through pragmatic decoding generally lowers quality and, somewhat counter-intuitively, increases repetitiveness in captions. Our mixed speaker, however, achieves a good balance between quality and informativeness.

The ability for variation in language use is necessary for speakers to achieve their conversational goals, for instance when referring to objects in visual environments. We argue that diversity should not be modelled as an independent objective in dialogue, but should rather be a result or by-product of goal-oriented language generation. Different lines of work in neural language generation investigated decoding methods for generating more diverse utterances, or increasing the informativity through pragmatic reasoning. We connect those lines of work and analyze how pragmatic reasoning during decoding affects the diversity of generated image captions. We find that boosting diversity itself does not result in more pragmatically informative captions, but pragmatic reasoning does increase lexical diversity. Finally, we discuss whether the gain in informativity is achieved in linguistically plausible ways.

Intuitive interaction with visual models becomes an increasingly important task in the field of Visualization (VIS) and verbal interaction represents a significant aspect of it. Vice versa, modeling verbal interaction in visual environments is a major trend in ongoing research in NLP. To date, research on Language & Vision, however, mostly happens at the intersection of NLP and Computer Vision (CV), and much less at the intersection of NLP and Visualization, which is an important area in Human-Computer Interaction (HCI). This paper presents a brief survey of recent work on interactive tasks and set-ups in NLP and Visualization. We discuss the respective methods, show interesting gaps, and conclude by suggesting neural, visually grounded dialogue modeling as a promising potential for NLIs for visual models.

We release ManyNames v2 (MN v2), a verified version of an object naming dataset that contains dozens of valid names per object for 25K images. We analyze issues in the data collection method originally employed, standard in Language & Vision (L&V), and find that the main source of noise in the data comes from simulating a naming context solely from an image with a target object marked with a bounding box, which causes subjects to sometimes disagree regarding which object is the target. We also find that both the degree of this uncertainty in the original data and the amount of true naming variation in MN v2 differs substantially across object domains. We use MN v2 to analyze a popular L&V model and demonstrate its effectiveness on the task of object naming. However, our fine-grained analysis reveals that what appears to be human-like model behavior is not stable across domains, e.g., the model confuses people and clothing objects much more frequently than humans do. We also find that standard evaluations underestimate the actual effectiveness of the naming model: on the single-label names of the original dataset (Visual Genome), it obtains −27% accuracy points than on MN v2, that includes all valid object names.

People choose particular names for objects, such as dog or puppy for a given dog. Object naming has been studied in Psycholinguistics, but has received relatively little attention in Computational Linguistics. We review resources from Language and Vision that could be used to study object naming on a large scale, discuss their shortcomings, and create a new dataset that affords more opportunities for analysis and modeling. Our dataset, ManyNames, provides 36 name annotations for each of 25K objects in images selected from VisualGenome. We highlight the challenges involved and provide a preliminary analysis of the ManyNames data, showing that there is a high level of agreement in naming, on average. At the same time, the average number of name types associated with an object is much higher in our dataset than in existing corpora for Language and Vision, such that ManyNames provides a rich resource for studying phenomena like hierarchical variation (chihuahua vs. dog), which has been discussed at length in the theoretical literature, and other less well studied phenomena like cross-classification (cake vs. dessert).

While certain types of instructions can be com-pactly expressed via images, there are situations where one might want to verbalise them, for example when directing someone. We investigate the task of Instruction Generation from Before/After Image Pairs which is to derive from images an instruction for effecting the implied change. For this, we make use of prior work on instruction following in a visual environment. We take an existing dataset, the BLOCKS data collected by Bisk et al. (2016) and investigate whether it is suitable for training an instruction generator as well. We find that it is, and investigate several simple baselines, taking these from the related task of image captioning. Through a series of experiments that simplify the task (by making image processing easier or completely side-stepping it; and by creating template-based targeted instructions), we investigate areas for improvement. We find that captioning models get some way towards solving the task, but have some difficulty with it, and future improvements must lie in the way the change is detected in the instruction.

In human cognition, world knowledge supports the perception of object colours: knowing that trees are typically green helps to perceive their colour in certain contexts. We go beyond previous studies on colour terms using isolated colour swatches and study visual grounding of colour terms in realistic objects. Our models integrate processing of visual information and object-specific knowledge via hard-coded (late) or learned (early) fusion. We find that both models consistently outperform a bottom-up baseline that predicts colour terms solely from visual inputs, but show interesting differences when predicting atypical colours of so-called colour diagnostic objects. Our models also achieve promising results when tested on new object categories not seen during training.

Zero-shot learning in Language & Vision is the task of correctly labelling (or naming) objects of novel categories. Another strand of work in L&V aims at pragmatically informative rather than “correct” object descriptions, e.g. in reference games. We combine these lines of research and model zero-shot reference games, where a speaker needs to successfully refer to a novel object in an image. Inspired by models of “rational speech acts”, we extend a neural generator to become a pragmatic speaker reasoning about uncertain object categories. As a result of this reasoning, the generator produces fewer nouns and names of distractor categories as compared to a literal speaker. We show that this conversational strategy for dealing with novel objects often improves communicative success, in terms of resolution accuracy of an automatic listener.

A lot of recent work in Language & Vision has looked at generating descriptions or referring expressions for objects in scenes of real-world images, though focusing mostly on relatively simple language like object names, color and location attributes (e.g., brown chair on the left). This paper presents work on Draw-and-Tell, a dataset of detailed descriptions for common objects in images where annotators have produced fine-grained attribute-centric expressions distinguishing a target object from a range of similar objects. Additionally, the dataset comes with hand-drawn sketches for each object. As Draw-and-Tell is medium-sized and contains a rich vocabulary, it constitutes an interesting challenge for CNN-LSTM architectures used in state-of-the-art image captioning models. We explore whether the additional modality given through sketches can help such a model to learn to accurately ground detailed language referring expressions to object shapes. Our results are encouraging.

We present a dataset consisting of what we call image description sequences, which are multi-sentence descriptions of the contents of an image. These descriptions were collected in a pseudo-interactive setting, where the describer was told to describe the given image to a listener who needs to identify the image within a set of images, and who successively asks for more information. As we show, this setup produced nicely structured data that, we think, will be useful for learning models capable of planning and realising such description discourses.

Image captioning models are typically trained on data that is collected from people who are asked to describe an image, without being given any further task context. As we argue here, this context independence is likely to cause problems for transferring to task settings in which image description is bound by task demands. We demonstrate that careful design of data collection is required to obtain image descriptions which are contextually bounded to a particular meta-level task. As a task, we use MeetUp!, a text-based communication game where two players have the goal of finding each other in a visual environment. To reach this goal, the players need to describe images representing their current location. We analyse a dataset from this domain and show that the nature of image descriptions found in MeetUp! is diverse, dynamic and rich with phenomena that are not present in descriptions obtained through a simple image captioning task, which we ran for comparison.

RNN-based sequence generation is now widely used in NLP and NLG (natural language generation). Most work focusses on how to train RNNs, even though also decoding is not necessarily straightforward: previous work on neural MT found seq2seq models to radically prefer short candidates, and has proposed a number of beam search heuristics to deal with this. In this work, we assess decoding strategies for referring expression generation with neural models. Here, expression length is crucial: output should neither contain too much or too little information, in order to be pragmatically adequate. We find that most beam search heuristics developed for MT do not generalize well to referring expression generation (REG), and do not generally outperform greedy decoding. We observe that beam search heuristics for termination seem to override the model’s knowledge of what a good stopping point is. Therefore, we also explore a recent approach called trainable decoding, which uses a small network to modify the RNN’s hidden state for better decoding results. We find this approach to consistently outperform greedy decoding for REG.

Modeling traditional NLG tasks with data-driven techniques has been a major focus of research in NLG in the past decade. We argue that existing modeling techniques are mostly tailored to textual data and are not sufficient to make NLG technology meet the requirements of agents which target fluid interaction and collaboration in the real world. We revisit interactive instruction giving as a challenge for datadriven NLG and, based on insights from previous GIVE challenges, propose that instruction giving should be addressed in a setting that involves visual grounding and spoken language. These basic design decisions will require NLG frameworks that are capable of monitoring their environment as well as timing and revising their verbal output. We believe that these are core capabilities for making NLG technology transferrable to interactive systems.

Corpora of referring expressions paired with their visual referents are a good source for learning word meanings directly grounded in visual representations. Here, we explore additional ways of extracting from them word representations linked to multi-modal context: through expressions that refer to the same object, and through expressions that refer to different objects in the same scene. We show that continuous meaning representations derived from these contexts capture complementary aspects of similarity, , even if not outperforming textual embeddings trained on very large amounts of raw text when tested on standard similarity benchmarks. We propose a new task for evaluating grounded meaning representations—detection of potentially co-referential phrases—and show that it requires precise denotational representations of attribute meanings, which our method provides.

There has recently been a lot of work trying to use images of referents of words for improving vector space meaning representations derived from text. We investigate the opposite direction, as it were, trying to improve visual word predictors that identify objects in images, by exploiting distributional similarity information during training. We show that for certain words (such as entry-level nouns or hypernyms), we can indeed learn better referential word meanings by taking into account their semantic similarity to other words. For other words, there is no or even a detrimental effect, compared to a learning setup that presents even semantically related objects as negative instances.

Current referring expression generation systems mostly deliver their output as one-shot, written expressions. We present on-going work on incremental generation of spoken expressions referring to objects in real-world images. This approach extends upon previous work using the words-as-classifier model for generation. We implement this generator in an incremental dialogue processing framework such that we can exploit an existing interface to incremental text-to-speech synthesis. Our system generates and synthesizes referring expressions while continuously observing non-verbal user reactions.

We propose a new shared task for tactical data-to-text generation in the domain of source code libraries. Specifically, we focus on text generation of function descriptions from example software projects. Data is drawn from existing resources used for studying the related problem of semantic parser induction, and spans a wide variety of both natural languages and programming languages. In this paper, we describe these existing resources, which will serve as training and development data for the task, and discuss plans for building new independent test sets.

A common convention in graphical user interfaces is to indicate a “wait state”, for example while a program is preparing a response, through a changed cursor state or a progress bar. What should the analogue be in a spoken conversational system? To address this question, we set up an experiment in which a human information provider (IP) was given their information only in a delayed and incremental manner, which systematically created situations where the IP had the turn but could not provide task-related information. Our data analysis shows that 1) IPs bridge the gap until they can provide information by re-purposing a whole variety of task- and grounding-related communicative actions (e.g. echoing the user’s request, signaling understanding, asserting partially relevant information), rather than being silent or explicitly asking for time (e.g. “please wait”), and that 2) IPs combined these actions productively to ensure an ongoing conversation. These results, we argue, indicate that natural conversational interfaces should also be able to manage their time flexibly using a variety of conversational resources.

We investigate object naming, which is an important sub-task of referring expression generation on real-world images. As opposed to mutually exclusive labels used in object recognition, object names are more flexible, subject to communicative preferences and semantically related to each other. Therefore, we investigate models of referential word meaning that link visual to lexical information which we assume to be given through distributional word embeddings. We present a model that learns individual predictors for object names that link visual and distributional aspects of word meaning during training. We show that this is particularly beneficial for zero-shot learning, as compared to projecting visual objects directly into the distributional space. In a standard object naming task, we find that different ways of combining lexical and visual information achieve very similar performance, though experiments on model combination suggest that they capture complementary aspects of referential meaning.

PentoRef is a corpus of task-oriented dialogues collected in systematically manipulated settings. The corpus is multilingual, with English and German sections, and overall comprises more than 20000 utterances. The dialogues are fully transcribed and annotated with referring expressions mapped to objects in corresponding visual scenes, which makes the corpus a rich resource for research on spoken referring expressions in generation and resolution. The corpus includes several sub-corpora that correspond to different dialogue situations where parameters related to interactivity, visual access, and verbal channel have been manipulated in systematic ways. The corpus thus lends itself to very targeted studies of reference in spontaneous dialogue.

In this paper, we investigate the usage of a non-canonical German passive alternation for ditransitive verbs, the recipient passive, in naturally occuring corpus data. We propose a classifier that predicts the voice of a ditransitive verb based on the contextually determined properties its arguments. As the recipient passive is a low frequent phenomenon, we first create a special data set focussing on German ditransitive verbs which are frequently used in the recipient passive. We use a broad-coverage grammar-based parser, the German LFG parser, to automatically annotate our data set for the morpho-syntactic properties of the involved predicate arguments. We train a Maximum Entropy classifier on the automatically annotated sentences and achieve an accuracy of 98.05%, clearly outperforming the baseline that always predicts active voice baseline (94.6%).

In this paper, we report on the design of a part-of-speech-tagset for Wolof and on the creation of a semi-automatically annotated gold standard. In order to achieve high-quality annotation relatively fast, we first generated an accurate lexicon that draws on existing word and name lists and takes into account inflectional and derivational morphology. The main motivation for the tagged corpus is to obtain data for training automatic taggers with machine learning approaches. Hence, we took machine learning considerations into account during tagset design and we present training experiments as part of this paper. The best automatic tagger achieves an accuracy of 95.2% in cross-validation experiments. We also wanted to create a basis for experimenting with annotation projection techniques, which exploit parallel corpora. For this reason, it was useful to use a part of the Bible as the gold standard corpus, for which sentence-aligned parallel versions in many languages are easy to obtain. We also report on preliminary experiments exploiting a statistical word alignment of the parallel text.