skip to main content

Explore Research Products in the NSF-PAR

Title: TAB-VCR: Tags and Attributes based VCR Baselines
Reasoning is an important ability that we learn from a very early age. Yet, reasoning is extremely hard for algorithms. Despite impressive recent progress that has been reported on tasks that necessitate reasoning, such as visual question answering and visual dialog, models often exploit biases in datasets. To develop models with better reasoning abilities, recently, the new visual commonsense reasoning (VCR) task has been introduced. Not only do models have to answer questions, but also do they have to provide a reason for the given answer. The proposed baseline achieved compelling results, leveraging a meticulously designed model composed of LSTM modules and attention nets. Here we show that a much simpler model obtained by ablating and pruning the existing intricate baseline can perform better with half the number of trainable parameters. By associating visual features with attribute information and better text to image grounding, we obtain further improvements for our simpler & effective baseline, TAB-VCR. We show that this approach results in a 5.3%, 4.4% and 6.5% absolute improvement over the previous state-of-the-art [103] on question answering, answer justification and holistic VCR.  more » « less
Award ID(s):
1725729
NSF-PAR ID:
10190038
Author(s) / Creator(s):
Date Published:
Journal Name:
Advances in Neural Information Processing Systems
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ( , ieee/cvf international conference on computer vision)
    Visual question answering (VQA) requires systems to perform concept-level reasoning by unifying unstructured (e.g., the context in question and answer; “QA context”) and structured (e.g., knowledge graph for the QA context and scene; “concept graph”) multimodal knowledge. Existing works typically combine a scene graph and a concept graph of the scene by connecting corresponding visual nodes and concept nodes, then incorporate the QA context representation to perform question answering. However, these methods only perform a unidirectional fusion from unstructured knowledge to structured knowledge, limiting their potential to capture joint reasoning over the heterogeneous modalities of knowledge. To perform more expressive reasoning, we propose VQA-GNN, a new VQA model that performs bidirectional fusion between unstructured and structured multimodal knowledge to obtain unified knowledge representations. Specifically, we inter-connect the scene graph and the concept graph through a super node that represents the QA context, and introduce a new multimodal GNN technique to perform inter-modal message passing for reasoning that mitigates representational gaps between modalities. On two challenging VQA tasks (VCR and GQA), our method outperforms strong baseline VQA methods by 3.2% on VCR (Q-AR) and 4.6% on GQA, suggesting its strength in performing concept-level reasoning. Ablation studies further demonstrate the efficacy of the bidirectional fusion and multimodal GNN method in unifying unstructured and structured multimodal knowledge. 
    more » « less
  2. ; ; ; ( , Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies)
    null (Ed.)
    Most existing research on visual question answering (VQA) is limited to information explicitly present in an image or a video. In this paper, we take visual understanding to a higher level where systems are challenged to answer questions that involve mentally simulating the hypothetical consequences of performing specific actions in a given scenario. Towards that end, we formulate a vision-language question answering task based on the CLEVR (Johnson et. al., 2017) dataset. We then modify the best existing VQA methods and propose baseline solvers for this task. Finally, we motivate the development of better vision-language models by providing insights about the capability of diverse architectures to perform joint reasoning over image-text modality. Our dataset setup scripts and codes will be made publicly available at https://github.com/shailaja183/clevr_hyp. 
    more » « less
  3. ; ; ; ; ; ; ( , Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume)
    Merlo, Paola ; Tiedemann, Jorg ; Tsarfaty, Reut (Ed.)
    GQA (CITATION) is a dataset for real-world visual reasoning and compositional question answering. We found that many answers predicted by the best vision-language models on the GQA dataset do not match the ground-truth answer but still are semantically meaningful and correct in the given context. In fact, this is the case with most existing visual question answering (VQA) datasets where they assume only one ground-truth answer for each question. We propose Alternative Answer Sets (AAS) of ground-truth answers to address this limitation, which is created automatically using off-the-shelf NLP tools. We introduce a semantic metric based on AAS and modify top VQA solvers to support multiple plausible answers for a question. We implement this approach on the GQA dataset and show the performance improvements. 
    more » « less
  4. ; ( , International Conference on Learning Representations)
    While large language models (LLMs) like GPT-3 have achieved impressive results on multiple choice question answering (MCQA) tasks in the zero, one, and few-shot settings, they generally lag behind the MCQA state of the art (SOTA). MCQA tasks have traditionally been presented to LLMs like cloze tasks. An LLM is conditioned on a question (without the associated answer options) and its chosen option is the one assigned the highest probability after normalization (for length, etc.). A more natural prompting approach is to present the question and answer options to the LLM jointly and have it output the symbol (e.g., “A”) associated with its chosen answer option. This approach allows the model to explicitly compare answer options, reduces computational costs, and mitigates the effects of tokenization scheme and answer option representations on answer selection. For the natural approach to be effective, the LLM it is used with must be able to associate answer options with the symbols that represent them. The LLM needs what we term multiple choice symbol binding (MCSB) ability. This ability varies greatly by model. We show that a model with high MCSB ability performs much better with the natural approach than with the traditional approach across 20 diverse datasets and largely closes the gap with the SOTA, suggesting that the MCQA ability of LLMs has been previously underestimated. 
    more » « less
  5. ; ; ; ; ; ( , GeoInformatica)
    Abstract

    Qualitative spatial/temporal reasoning (QSR/QTR) plays a key role in research on human cognition, e.g., as it relates to navigation, as well as in work on robotics and artificial intelligence. Although previous work has mainly focused on various spatial and temporal calculi, more recently representation learning techniques such as embedding have been applied to reasoning and inference tasks such as query answering and knowledge base completion. These subsymbolic and learnable representations are well suited for handling noise and efficiency problems that plagued prior work. However, applying embedding techniques to spatial and temporal reasoning has received little attention to date. In this paper, we explore two research questions: (1) How do embedding-based methods perform empirically compared to traditional reasoning methods on QSR/QTR problems? (2) If the embedding-based methods are better, what causes this superiority? In order to answer these questions, we first propose a hyperbolic embedding model, called HyperQuaternionE, to capture varying properties of relations (such as symmetry and anti-symmetry), to learn inversion relations and relation compositions (i.e., composition tables), and to model hierarchical structures over entities induced by transitive relations. We conduct various experiments on two synthetic datasets to demonstrate the advantages of our proposed embedding-based method against existing embedding models as well as traditional reasoners with respect to entity inference and relation inference. Additionally, our qualitative analysis reveals that our method is able to learn conceptual neighborhoods implicitly. We conclude that the success of our method is attributed to its ability to model composition tables and learn conceptual neighbors, which are among the core building blocks of QSR/QTR.

     
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email