Iacer Calixto


Multi3Generation: Multitask, Multilingual, Multimodal Language Generation
Anabela Barreiro | José GC de Souza | Albert Gatt | Mehul Bhatt | Elena Lloret | Aykut Erdem | Dimitra Gkatzia | Helena Moniz | Irene Russo | Fabio Kepler | Iacer Calixto | Marcin Paprzycki | François Portet | Isabelle Augenstein | Mirela Alhasani
Proceedings of the 23rd Annual Conference of the European Association for Machine Translation

This paper presents the Multitask, Multilingual, Multimodal Language Generation COST Action – Multi3Generation (CA18231), an interdisciplinary network of research groups working on different aspects of language generation. This “meta-paper” will serve as reference for citations of the Action in future publications. It presents the objectives, challenges and a the links for the achieved outcomes.

VALSE: A Task-Independent Benchmark for Vision and Language Models Centered on Linguistic Phenomena
Letitia Parcalabescu | Michele Cafagna | Lilitta Muradjan | Anette Frank | Iacer Calixto | Albert Gatt
Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

We propose VALSE (Vision And Language Structured Evaluation), a novel benchmark designed for testing general-purpose pretrained vision and language (V&L) models for their visio-linguistic grounding capabilities on specific linguistic phenomena. VALSE offers a suite of six tests covering various linguistic constructs. Solving these requires models to ground linguistic phenomena in the visual modality, allowing more fine-grained evaluations than hitherto possible. We build VALSE using methods that support the construction of valid foils, and report results from evaluating five widely-used V&L models. Our experiments suggest that current models have considerable difficulty addressing most phenomena. Hence, we expect VALSE to serve as an important benchmark to measure future progress of pretrained V&L models from a linguistic perspective, complementing the canonical task-centred V&L evaluations.


Wikipedia Entities as Rendezvous across Languages: Grounding Multilingual Language Models by Predicting Wikipedia Hyperlinks
Iacer Calixto | Alessandro Raganato | Tommaso Pasini
Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies

Masked language models have quickly become the de facto standard when processing text. Recently, several approaches have been proposed to further enrich word representations with external knowledge sources such as knowledge graphs. However, these models are devised and evaluated in a monolingual setting only. In this work, we propose a language-independent entity prediction task as an intermediate training procedure to ground word representations on entity semantics and bridge the gap across different languages by means of a shared vocabulary of entities. We show that our approach effectively injects new lexical-semantic knowledge into neural models, improving their performance on different semantic tasks in the zero-shot crosslingual setting. As an additional advantage, our intermediate training does not require any supplementary input, allowing our models to be applied to new datasets right away. In our experiments, we use Wikipedia articles in up to 100 languages and already observe consistent gains compared to strong baselines when predicting entities using only the English Wikipedia. Further adding extra languages lead to improvements in most tasks up to a certain point, but overall we found it non-trivial to scale improvements in model transferability by training on ever increasing amounts of Wikipedia languages.

VisualSem: a high-quality knowledge graph for vision and language
Houda Alberts | Ningyuan Huang | Yash Deshpande | Yibo Liu | Kyunghyun Cho | Clara Vania | Iacer Calixto
Proceedings of the 1st Workshop on Multilingual Representation Learning

An exciting frontier in natural language understanding (NLU) and generation (NLG) calls for (vision-and-) language models that can efficiently access external structured knowledge repositories. However, many existing knowledge bases only cover limited domains, or suffer from noisy data, and most of all are typically hard to integrate into neural language pipelines. To fill this gap, we release VisualSem: a high-quality knowledge graph (KG) which includes nodes with multilingual glosses, multiple illustrative images, and visually relevant relations. We also release a neural multi-modal retrieval model that can use images or sentences as inputs and retrieves entities in the KG. This multi-modal retrieval model can be integrated into any (neural network) model pipeline. We encourage the research community to use VisualSem for data augmentation and/or as a source of grounding, among other possible uses. VisualSem as well as the multi-modal retrieval models are publicly available and can be downloaded in this URL: https://github.com/iacercalixto/visualsem.

Seeing past words: Testing the cross-modal capabilities of pretrained V&L models on counting tasks
Letitia Parcalabescu | Albert Gatt | Anette Frank | Iacer Calixto
Proceedings of the 1st Workshop on Multimodal Semantic Representations (MMSR)

We investigate the reasoning ability of pretrained vision and language (V&L) models in two tasks that require multimodal integration: (1) discriminating a correct image-sentence pair from an incorrect one, and (2) counting entities in an image. We evaluate three pretrained V&L models on these tasks: ViLBERT, ViLBERT 12-in-1 and LXMERT, in zero-shot and finetuned settings. Our results show that models solve task (1) very well, as expected, since all models are pretrained on task (1). However, none of the pretrained V&L models is able to adequately solve task (2), our counting probe, and they cannot generalise to out-of-distribution quantities. We propose a number of explanations for these findings: LXMERT (and to some extent ViLBERT 12-in-1) show some evidence of catastrophic forgetting on task (1). Concerning our results on the counting probe, we find evidence that all models are impacted by dataset bias, and also fail to individuate entities in the visual input. While a selling point of pretrained V&L models is their ability to solve complex tasks, our findings suggest that understanding their reasoning and grounding capabilities requires more targeted investigations on specific phenomena.

Proceedings of the 6th Workshop on Representation Learning for NLP (RepL4NLP-2021)
Anna Rogers | Iacer Calixto | Ivan Vulić | Naomi Saphra | Nora Kassner | Oana-Maria Camburu | Trapit Bansal | Vered Shwartz
Proceedings of the 6th Workshop on Representation Learning for NLP (RepL4NLP-2021)


Are Scene Graphs Good Enough to Improve Image Captioning?
Victor Milewski | Marie-Francine Moens | Iacer Calixto
Proceedings of the 1st Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 10th International Joint Conference on Natural Language Processing

Many top-performing image captioning models rely solely on object features computed with an object detection model to generate image descriptions. However, recent studies propose to directly use scene graphs to introduce information about object relations into captioning, hoping to better describe interactions between objects. In this work, we thoroughly investigate the use of scene graphs in image captioning. We empirically study whether using additional scene graph encoders can lead to better image descriptions and propose a conditional graph attention network (C-GAT), where the image captioning decoder state is used to condition the graph updates. Finally, we determine to what extent noise in the predicted scene graphs influence caption quality. Overall, we find no significant difference between models that use scene graph features and models that only use object detection features across different captioning metrics, which suggests that existing scene graph generation models are still too noisy to be useful in image captioning. Moreover, although the quality of predicted scene graphs is very low in general, when using high quality scene graphs we obtain gains of up to 3.3 CIDEr compared to a strong Bottom-Up Top-Down baseline.

English Intermediate-Task Training Improves Zero-Shot Cross-Lingual Transfer Too
Jason Phang | Iacer Calixto | Phu Mon Htut | Yada Pruksachatkun | Haokun Liu | Clara Vania | Katharina Kann | Samuel R. Bowman
Proceedings of the 1st Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 10th International Joint Conference on Natural Language Processing

Intermediate-task training—fine-tuning a pretrained model on an intermediate task before fine-tuning again on the target task—often improves model performance substantially on language understanding tasks in monolingual English settings. We investigate whether English intermediate-task training is still helpful on non-English target tasks. Using nine intermediate language-understanding tasks, we evaluate intermediate-task transfer in a zero-shot cross-lingual setting on the XTREME benchmark. We see large improvements from intermediate training on the BUCC and Tatoeba sentence retrieval tasks and moderate improvements on question-answering target tasks. MNLI, SQuAD and HellaSwag achieve the best overall results as intermediate tasks, while multi-task intermediate offers small additional improvements. Using our best intermediate-task models for each target task, we obtain a 5.4 point improvement over XLM-R Large on the XTREME benchmark, setting the state of the art as of June 2020. We also investigate continuing multilingual MLM during intermediate-task training and using machine-translated intermediate-task data, but neither consistently outperforms simply performing English intermediate-task training.

Can Wikipedia Categories Improve Masked Language Model Pretraining?
Diksha Meghwal | Katharina Kann | Iacer Calixto | Stanislaw Jastrzebski
Proceedings of the The Fourth Widening Natural Language Processing Workshop

Pretrained language models have obtained impressive results for a large set of natural language understanding tasks. However, training these models is computationally expensive and requires huge amounts of data. Thus, it would be desirable to automatically detect groups of more or less important examples. Here, we investigate if we can leverage sources of information which are commonly overlooked, Wikipedia categories as listed in DBPedia, to identify useful or harmful data points during pretraining. We define an experimental setup in which we analyze correlations between language model perplexity on specific clusters and downstream NLP task performances during pretraining. Our experiments show that Wikipedia categories are not a good indicator of the importance of specific sentences for pretraining.


Latent Variable Model for Multi-modal Translation
Iacer Calixto | Miguel Rios | Wilker Aziz
Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics

In this work, we propose to model the interaction between visual and textual features for multi-modal neural machine translation (MMT) through a latent variable model. This latent variable can be seen as a multi-modal stochastic embedding of an image and its description in a foreign language. It is used in a target-language decoder and also to predict image features. Importantly, our model formulation utilises visual and textual inputs during training but does not require that images be available at test time. We show that our latent variable MMT formulation improves considerably over strong baselines, including a multi-task learning approach (Elliott and Kadar, 2017) and a conditional variational auto-encoder approach (Toyama et al., 2016). Finally, we show improvements due to (i) predicting image features in addition to only conditioning on them, (ii) imposing a constraint on the KL term to promote models with non-negligible mutual information between inputs and latent variable, and (iii) by training on additional target-language image descriptions (i.e. synthetic data).


Doubly-Attentive Decoder for Multi-modal Neural Machine Translation
Iacer Calixto | Qun Liu | Nick Campbell
Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)

We introduce a Multi-modal Neural Machine Translation model in which a doubly-attentive decoder naturally incorporates spatial visual features obtained using pre-trained convolutional neural networks, bridging the gap between image description and translation. Our decoder learns to attend to source-language words and parts of an image independently by means of two separate attention mechanisms as it generates words in the target language. We find that our model can efficiently exploit not just back-translated in-domain multi-modal data but also large general-domain text-only MT corpora. We also report state-of-the-art results on the Multi30k data set.

Sentence-Level Multilingual Multi-modal Embedding for Natural Language Processing
Iacer Calixto | Qun Liu
Proceedings of the International Conference Recent Advances in Natural Language Processing, RANLP 2017

We propose a novel discriminative ranking model that learns embeddings from multilingual and multi-modal data, meaning that our model can take advantage of images and descriptions in multiple languages to improve embedding quality. To that end, we introduce an objective function that uses pairwise ranking adapted to the case of three or more input sources. We compare our model against different baselines, and evaluate the robustness of our embeddings on image–sentence ranking (ISR), semantic textual similarity (STS), and neural machine translation (NMT). We find that the additional multilingual signals lead to improvements on all three tasks, and we highlight that our model can be used to consistently improve the adequacy of translations generated with NMT models when re-ranking n-best lists.

Incorporating Global Visual Features into Attention-based Neural Machine Translation.
Iacer Calixto | Qun Liu
Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing

We introduce multi-modal, attention-based neural machine translation (NMT) models which incorporate visual features into different parts of both the encoder and the decoder. Global image features are extracted using a pre-trained convolutional neural network and are incorporated (i) as words in the source sentence, (ii) to initialise the encoder hidden state, and (iii) as additional data to initialise the decoder hidden state. In our experiments, we evaluate translations into English and German, how different strategies to incorporate global image features compare and which ones perform best. We also study the impact that adding synthetic multi-modal, multilingual data brings and find that the additional data have a positive impact on multi-modal NMT models. We report new state-of-the-art results and our best models also significantly improve on a comparable phrase-based Statistical MT (PBSMT) model trained on the Multi30k data set according to all metrics evaluated. To the best of our knowledge, it is the first time a purely neural model significantly improves over a PBSMT model on all metrics evaluated on this data set.

Using Images to Improve Machine-Translating E-Commerce Product Listings.
Iacer Calixto | Daniel Stein | Evgeny Matusov | Pintu Lohar | Sheila Castilho | Andy Way
Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers

In this paper we study the impact of using images to machine-translate user-generated e-commerce product listings. We study how a multi-modal Neural Machine Translation (NMT) model compares to two text-only approaches: a conventional state-of-the-art attentional NMT and a Statistical Machine Translation (SMT) model. User-generated product listings often do not constitute grammatical or well-formed sentences. More often than not, they consist of the juxtaposition of short phrases or keywords. We train our models end-to-end as well as use text-only and multi-modal NMT models for re-ranking n-best lists generated by an SMT model. We qualitatively evaluate our user-generated training data also analyse how adding synthetic data impacts the results. We evaluate our models quantitatively using BLEU and TER and find that (i) additional synthetic data has a general positive impact on text-only and multi-modal NMT models, and that (ii) using a multi-modal NMT model for re-ranking n-best lists improves TER significantly across different n-best list sizes.

Human Evaluation of Multi-modal Neural Machine Translation: A Case-Study on E-Commerce Listing Titles
Iacer Calixto | Daniel Stein | Evgeny Matusov | Sheila Castilho | Andy Way
Proceedings of the Sixth Workshop on Vision and Language

In this paper, we study how humans perceive the use of images as an additional knowledge source to machine-translate user-generated product listings in an e-commerce company. We conduct a human evaluation where we assess how a multi-modal neural machine translation (NMT) model compares to two text-only approaches: a conventional state-of-the-art attention-based NMT and a phrase-based statistical machine translation (PBSMT) model. We evaluate translations obtained with different systems and also discuss the data set of user-generated product listings, which in our case comprises both product listings and associated images. We found that humans preferred translations obtained with a PBSMT system to both text-only and multi-modal NMT over 56% of the time. Nonetheless, human evaluators ranked translations from a multi-modal NMT model as better than those of a text-only NMT over 88% of the time, which suggests that images do help NMT in this use-case.

Linguistic realisation as machine translation: Comparing different MT models for AMR-to-text generation
Thiago Castro Ferreira | Iacer Calixto | Sander Wubben | Emiel Krahmer
Proceedings of the 10th International Conference on Natural Language Generation

In this paper, we study AMR-to-text generation, framing it as a translation task and comparing two different MT approaches (Phrase-based and Neural MT). We systematically study the effects of 3 AMR preprocessing steps (Delexicalisation, Compression, and Linearisation) applied before the MT phase. Our results show that preprocessing indeed helps, although the benefits differ for the two MT models.

DCU System Report on the WMT 2017 Multi-modal Machine Translation Task
Iacer Calixto | Koel Dutta Chowdhury | Qun Liu
Proceedings of the Second Conference on Machine Translation


DCU-UvA Multimodal MT System Report
Iacer Calixto | Desmond Elliott | Stella Frank
Proceedings of the First Conference on Machine Translation: Volume 2, Shared Task Papers

Developing a Dataset for Evaluating Approaches for Document Expansion with Images
Debasis Ganguly | Iacer Calixto | Gareth Jones
Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC'16)

Motivated by the adage that a “picture is worth a thousand words” it can be reasoned that automatically enriching the textual content of a document with relevant images can increase the readability of a document. Moreover, features extracted from the additional image data inserted into the textual content of a document may, in principle, be also be used by a retrieval engine to better match the topic of a document with that of a given query. In this paper, we describe our approach of building a ground truth dataset to enable further research into automatic addition of relevant images to text documents. The dataset is comprised of the official ImageCLEF 2010 collection (a collection of images with textual metadata) to serve as the images available for automatic enrichment of text, a set of 25 benchmark documents that are to be enriched, which in this case are children’s short stories, and a set of manually judged relevant images for each query story obtained by the standard procedure of depth pooling. We use this benchmark dataset to evaluate the effectiveness of standard information retrieval methods as simple baselines for this task. The results indicate that using the whole story as a weighted query, where the weight of each query term is its tf-idf value, achieves an precision of 0:1714 within the top 5 retrieved images on an average.


Automatic Text Simplification for Spanish: Comparative Evaluation of Various Simplification Strategies
Sanja Štajner | Iacer Calixto | Horacio Saggion
Proceedings of the International Conference Recent Advances in Natural Language Processing


Target-Centric Features for Translation Quality Estimation
Chris Hokamp | Iacer Calixto | Joachim Wagner | Jian Zhang
Proceedings of the Ninth Workshop on Statistical Machine Translation