Recent computational models of the acquisition of spoken language via grounding in perception exploit associations between spoken and visual modalities and learn to represent speech and visual data in a joint vector space. A major unresolved issue from the point of ecological validity is the training data, typically consisting of images or videos paired with spoken descriptions of what is depicted. Such a setup guarantees an unrealistically strong correlation between speech and the visual data. In the real world the coupling between the linguistic and the visual modality is loose, and often confounded by correlations with non-semantic aspects of the speech signal. Here we address this shortcoming by using a dataset based on the children’s cartoon Peppa Pig. We train a simple bi-modal architecture on the portion of the data consisting of dialog between characters, and evaluate on segments containing descriptive narrations. Despite the weak and confounded signal in this training data, our model succeeds at learning aspects of the visual semantics of spoken language.

Recent advances in vision-and-language modeling have seen the development of Transformer architectures that achieve remarkable performance on multimodal reasoning tasks. Yet, the exact capabilities of these black-box models are still poorly understood. While much of previous work has focused on studying their ability to learn meaning at the word-level, their ability to track syntactic dependencies between words has received less attention. We take a first step in closing this gap by creating a new multimodal task targeted at evaluating understanding of predicate-noun dependencies in a controlled setup. We evaluate a range of state-of-the-art models and find that their performance on the task varies considerably, with some models performing relatively well and others at chance level. In an effort to explain this variability, our analyses indicate that the quality (and not only sheer quantity) of pretraining data is essential. Additionally, the best performing models leverage fine-grained multimodal pretraining objectives in addition to the standard image-text matching objectives. This study highlights that targeted and controlled evaluations are a crucial step for a precise and rigorous test of the multimodal knowledge of vision-and-language models.

Children learn the meaning of words and sentences in their native language at an impressive speed and from highly ambiguous input. To account for this learning, previous computational modeling has focused mainly on the study of perception-based mechanisms like cross-situational learning. However, children do not learn only by exposure to the input. As soon as they start to talk, they practice their knowledge in social interactions and they receive feedback from their caregivers. In this work, we propose a model integrating both perception- and production-based learning using artificial neural networks which we train on a large corpus of crowd-sourced images with corresponding descriptions. We found that production-based learning improves performance above and beyond perception-based learning across a wide range of semantic tasks including both word- and sentence-level semantics. In addition, we documented a synergy between these two mechanisms, where their alternation allows the model to converge on more balanced semantic knowledge. The broader impact of this work is to highlight the importance of modeling language learning in the context of social interactions where children are not only understood as passively absorbing the input, but also as actively participating in the construction of their linguistic knowledge.

When learning their native language, children acquire the meanings of words and sentences from highly ambiguous input without much explicit supervision. One possible learning mechanism is cross-situational learning, which has been successfully tested in laboratory experiments with children. Here we use Artificial Neural Networks to test if this mechanism scales up to more natural language and visual scenes using a large dataset of crowd-sourced images with corresponding descriptions. We evaluate learning using a series of tasks inspired by methods commonly used in laboratory studies of language acquisition. We show that the model acquires rich semantic knowledge both at the word- and sentence-level, mirroring the patterns and trajectory of learning in early childhood. Our work highlights the usefulness of low-level co-occurrence statistics across modalities in facilitating the early acquisition of higher-level semantic knowledge.

Image captioning models are usually evaluated on their ability to describe a held-out set of images, not on their ability to generalize to unseen concepts. We study the problem of compositional generalization, which measures how well a model composes unseen combinations of concepts when describing images. State-of-the-art image captioning models show poor generalization performance on this task. We propose a multi-task model to address the poor performance, that combines caption generation and image–sentence ranking, and uses a decoding mechanism that re-ranks the captions according their similarity to the image. This model is substantially better at generalizing to unseen combinations of concepts compared to state-of-the-art captioning models.

We present a detailed comparison of two types of sequence to sequence models trained to conduct a compositional task. The models are architecturally identical at inference time, but differ in the way that they are trained: our baseline model is trained with a task-success signal only, while the other model receives additional supervision on its attention mechanism (Attentive Guidance), which has shown to be an effective method for encouraging more compositional solutions. We first confirm that the models with attentive guidance indeed infer more compositional solutions than the baseline, by training them on the lookup table task presented by Liska et al. (2019). We then do an in-depth analysis of the structural differences between the two model types, focusing in particular on the organisation of the parameter space and the hidden layer activations and find noticeable differences in both these aspects. Guided networks focus more on the components of the input rather than the sequence as a whole and develop small functional groups of neurons with specific purposes that use their gates more selectively. Results from parameter heat maps, component swapping and graph analysis also indicate that guided networks exhibit a more modular structure with a small number of specialized, strongly connected neurons.