Visual storytelling systems generate multi-sentence stories from image sequences. In this task, capturing contextual information and bridging visual variation bring additional challenges. We propose a simple yet effective framework that leverages the generalization capabilities of pretrained foundation models, only training a lightweight vision-language mapping network to connect modalities, while incorporating context to enhance coherence. We introduce a multimodal contrastive objective that also improves visual relevance and story informativeness. Extensive experimental results, across both automatic metrics and human evaluations, demonstrate that the stories generated by our framework are diverse, coherent, informative, and interesting.

Grounding has been argued to be a crucial component towards the development of more complete and truly semantically competent artificial intelligence systems. Literature has divided into two camps: While some argue that grounding allows for qualitatively different generalizations, others believe it can be compensated by mono-modal data quantity. Limited empirical evidence has emerged for or against either position, which we argue is due to the methodological challenges that come with studying grounding and its effects on NLP systems. In this paper, we establish a methodological framework for studying what the effects are—if any—of providing models with richer input sources than text-only. The crux of it lies in the construction of comparable samples of populations of models trained on different input modalities, so that we can tease apart the qualitative effects of different input sources from quantifiable model performances. Experiments using this framework reveal qualitative differences in model behavior between cross-modally grounded, cross-lingually grounded, and ungrounded models, which we measure both at a global dataset level as well as for specific word representations, depending on how concrete their semantics is.

Multimodal embeddings aim to enrich the semantic information in neural representations of language compared to text-only models. While different embeddings exhibit different applicability and performance on downstream tasks, little is known about the systematic representation differences attributed to the visual modality. Our paper compares word embeddings from three vision-and-language models (CLIP, OpenCLIP and Multilingual CLIP, Radford et al. 2021; Ilharco et al. 2021; Carlsson et al. 2022) and three text-only models, with static (FastText, Bojanowski et al. 2017) as well as contextual representations (multilingual BERT Devlin et al. 2018; XLM-RoBERTa, Conneau et al. 2019). This is the first large-scale study of the effect of visual grounding on language representations, including 46 semantic parameters. We identify meaning properties and relations that characterize words whose embeddings are most affected by the inclusion of visual modality in the training data; that is, points where visual grounding turns out most important. We find that the effect of visual modality correlates most with denotational semantic properties related to concreteness, but is also detected for several specific semantic classes, as well as for valence, a sentiment-related connotational property of linguistic expressions.

How does the word analogy task fit in the modern NLP landscape? Given the rarity of comparable multilingual benchmarks and the lack of a consensual evaluation protocol for contextual models, this remains an open question. In this paper, we introduce MATS: a multilingual analogy dataset, covering forty analogical relations in six languages, and evaluate human as well as static and contextual embedding performances on the task. We find that not all analogical relations are equally straightforward for humans, static models remain competitive with contextual embeddings, and optimal settings vary across languages and analogical relations. Several key challenges remain, including creating benchmarks that align with human reasoning and understanding what drives differences across methodologies.

Can recurrent neural nets, inspired by human sequential data processing, learn to understand language? We construct simplified data sets reflecting core properties of natural language as modeled in formal syntax and semantics: recursive syntactic structure and compositionality. We find LSTM and GRU networks to generalize to compositional interpretation well, but only in the most favorable learning settings, with a well-paced curriculum, extensive training data, and left-to-right (but not right-to-left) composition.

Word embeddings have advanced the state of the art in NLP across numerous tasks. Understanding the contents of dense neural representations is of utmost interest to the computational semantics community. We propose to focus on relating these opaque word vectors with human-readable definitions, as found in dictionaries This problem naturally divides into two subtasks: converting definitions into embeddings, and converting embeddings into definitions. This task was conducted in a multilingual setting, using comparable sets of embeddings trained homogeneously.

Pretrained embeddings based on the Transformer architecture have taken the NLP community by storm. We show that they can mathematically be reframed as a sum of vector factors and showcase how to use this reframing to study the impact of each component. We provide evidence that multi-head attentions and feed-forwards are not equally useful in all downstream applications, as well as a quantitative overview of the effects of finetuning on the overall embedding space. This approach allows us to draw connections to a wide range of previous studies, from vector space anisotropy to attention weights.

Standard image caption generation systems produce generic descriptions of images and do not utilize any contextual information or world knowledge. In particular, they are unable to generate captions that contain references to the geographic context of an image, for example, the location where a photograph is taken or relevant geographic objects around an image location. In this paper, we develop a geo-aware image caption generation system, which incorporates geographic contextual information into a standard image captioning pipeline. We propose a way to build an image-specific representation of the geographic context and adapt the caption generation network to produce appropriate geographic names in the image descriptions. We evaluate our system on a novel captioning dataset that contains contextualized captions and geographic metadata and achieve substantial improvements in BLEU, ROUGE, METEOR and CIDEr scores. We also introduce a new metric to assess generated geographic references directly and empirically demonstrate our system’s ability to produce captions with relevant and factually accurate geographic referencing.

Compositionality is a widely discussed property of natural languages, although its exact definition has been elusive. We focus on the proposal that compositionality can be assessed by measuring meaning-form correlation. We analyze meaning-form correlation on three sets of languages: (i) artificial toy languages tailored to be compositional, (ii) a set of English dictionary definitions, and (iii) a set of English sentences drawn from literature. We find that linguistic phenomena such as synonymy and ungrounded stop-words weigh on MFC measurements, and that straightforward methods to mitigate their effects have widely varying results depending on the dataset they are applied to. Data and code are made publicly available.

La génération de définitions est une tâche récente qui vise à produire des définitions lexicographiques à partir de plongements lexicaux. Nous remarquons deux lacunes : (i) l’état de l’art actuel ne s’est penché que sur l’anglais et le chinois, et (ii) l’utilisation escomptée en tant que méthode d’évaluation des plongements lexicaux doit encore être vérifiée. Pour y remédier, nous proposons un jeu de données pour la génération de définitions en français, ainsi qu’une évaluation des performances d’un modèle de génération de définitions simple selon les plongements lexicaux fournis en entrée.

This paper presents our work in long and short form choice, a significant question of lexical choice, which plays an important role in many Natural Language Understanding tasks. Long and short form sharing at least one identical word meaning but with different number of syllables is a highly frequent linguistic phenomenon in Chinese like 老虎-虎(laohu-hu, tiger)

Defining words in a textual context is a useful task both for practical purposes and for gaining insight into distributed word representations. Building on the distributional hypothesis, we argue here that the most natural formalization of definition modeling is to treat it as a sequence-to-sequence task, rather than a word-to-sequence task: given an input sequence with a highlighted word, generate a contextually appropriate definition for it. We implement this approach in a Transformer-based sequence-to-sequence model. Our proposal allows to train contextualization and definition generation in an end-to-end fashion, which is a conceptual improvement over earlier works. We achieve state-of-the-art results both in contextual and non-contextual definition modeling.

Between 80% and 90% of all Chinese words have long and short form such as 老虎/虎 (lao-hu/hu , tiger) (Duanmu:2013). Consequently, the choice between long and short forms is a key problem for lexical choice across NLP and NLG. Following an earlier work on abbreviations in English (Mahowald et al, 2013), we bring a probabilistic perspective to these questions, using both a behavioral and a corpus-based approach. We hypothesized that there is a higher probability of choosing short form in supportive context than in neutral context in Mandarin. Consistent with our prediction, our findings revealed that predictability of contexts makes effect on speakers’ long and short form choice.

In this submission I report work in progress on learning simplified interpreted languages by means of recurrent models. The data is constructed to reflect core properties of natural language as modeled in formal syntax and semantics. Preliminary results suggest that LSTM networks do generalise to compositional interpretation, albeit only in the most favorable learning setting.

This paper describes the SemEval 2018 Task 10 on Capturing Discriminative Attributes. Participants were asked to identify whether an attribute could help discriminate between two concepts. For example, a successful system should determine that ‘urine’ is a discriminating feature in the word pair ‘kidney’, ‘bone’. The aim of the task is to better evaluate the capabilities of state of the art semantic models, beyond pure semantic similarity. The task attracted submissions from 21 teams, and the best system achieved a 0.75 F1 score.

In this paper we present a novel application of compositional distributional semantic models (CDSMs): prediction of lexical typology. The paper introduces the notion of typological closeness, which is a novel rigorous formalization of semantic similarity based on comparison of multilingual data. Starting from the Moscow Database of Qualitative Features for adjective typology, we create four datasets of typological closeness, on which we test a range of distributional semantic models. We show that, on the one hand, vector representations of phrases based on data from one language can be used to predict how words within the phrase translate into different languages, and, on the other hand, that typological data can serve as a semantic benchmark for distributional models. We find that compositional distributional models, especially parametric ones, perform way above non-compositional alternatives on the task.

Corpus-based distributional semantic models capture degrees of semantic relatedness among the words of very large vocabularies, but have problems with logical phenomena such as entailment, that are instead elegantly handled by model-theoretic approaches, which, in turn, do not scale up. We combine the advantages of the two views by inducing a mapping from distributional vectors of words (or sentences) into a Boolean structure of the kind in which natural language terms are assumed to denote. We evaluate this Boolean Distributional Semantic Model (BDSM) on recognizing entailment between words and sentences. The method achieves results comparable to a state-of-the-art SVM, degrades more gracefully when less training data are available and displays interesting qualitative properties.