Generating Visual Explanations

By Lisa Anne Hendricks, Zeynep Akata, Marcus Rohrbach, Jeff Donahue, Bernt Schiele, Trevor Darrell. 2016

658 citations recorded by the source corpus.

Clearly explaining a rationale for a classification decision to an end user can be as important as the decision itself. Existing approaches for deep visual recognition are generally opaque and do not output any justification text; contemporary vision-language models can describe image content but fail to take into account class-discriminative image aspects which justify visual predictions. We propose a new model that focuses on the discriminating properties of the visible object, jointly predicts a class label, and explains why the predicted label is appropriate for the image. Through a novel loss function based on sampling and reinforcement learning, our model learns to generate sentences that realize a global sentence property, such as class specificity. Our results on the CUB dataset show that our model is able to generate explanations which are not only consistent with an image but also more discriminative than descriptions produced by existing captioning methods.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Measuring and Improving BERT’s Mathematical Abilities by Predicting the Order of Reasoning. (2021, background): “Systems that produce explanations either in the form of the language (Ling et al., 2017; Hendricks et al., 2016), attention (Bahdanau et al., 2014; Mnih et al., 2014; Gulcehre et al., 2016; Malinowski et al., 2018; Xu and Saenko, 2016; Yang et al., 2016), phrase selection (Lei et al., 2016),…”
Program Induction by Rationale Generation: Learning to Solve and Explain Algebraic Word Problems (2017, background): “This work is thus related to models that can explain or rationalize their decisions (Hendricks et al., 2016; Harrison et al., 2017).”
ALICE: Active Learning with Contrastive Natural Language Explanations (2020, background): “Distantly related to our work, Hendricks et al. (2016) propose to generate explanations for image classifiers but they do not explore improving the classifiers with the explanations.”
Shaping Visual Representations with Language for Few-Shot Classification (2019, neutral): “…which generate natural language explanations of decisions for interpretability for both textual (e.g. natural language inference; [3]) and visual [17, 18] tasks, but here we examine whether this act of predicting language can actually improve downstream task performance; similar ideas have been…”
Can language models learn from explanations in context? (2022, methodology): “…(e.g. Wei et al., 2022; Reynolds and McDonell, 2021), as well as a huge variety of broader work beyond prompting on topics like training or tuning with explanations (e.g. Camburu et al., 2018; Ling et al., 2017; Hendricks et al., 2016; Narang et al., 2020; Lertvittayakumjorn and Toni, 2021).”
Reframing Human-AI Collaboration for Generating Free-Text Explanations (2021, methodology): “These free-text explanations provide a natural interface between machine computation and human end-users, in contrast to explanations limited to highlighting inputs or predicting numerical forms (Hendricks et al., 2016; Camburu et al., 2018).”
Few-Shot Out-of-Domain Transfer Learning of Natural Language Explanations (2021, background): “…language explanations (NLEs) for the ground-truth labels at training time and generate such explanations for their decisions at deployment time (Hendricks et al., 2016; Ling et al., 2017; Park et al., 2018; Camburu et al., 2018; Kim et al., 2018; Rajani et al., 2019; Camburu et al., 2020;…”
Measuring Association Between Labels and Free-Text Rationales (2020, background): “This is the most common approach to free-text rationalization (Hendricks et al., 2016; Kim et al., 2018; Camburu et al., 2018; Ehsan et al., 2018; Liu et al., 2019; Wu and Mooney, 2019; Narang et al., 2020; Do et al., 2020), but little is understood about its faithfulness.”
Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering (2016, background): “[17] generates a natural language explanation (sentence) for image categories.”
From Recognition to Cognition: Visual Commonsense Reasoning (2018, background): “This has motivated work in having models provide explanations for their behavior, in the form of a natural language sentence [31, 9, 41] or an attention map [32, 35, 37].”
Explaining Explanations: An Overview of Interpretability of Machine Learning (2018, background): “Explanation generation has been demonstrated as part of systems for visual question answering [71] as well as in ﬁne-grained image classiﬁcation [72].”
Neural-Symbolic Computing: An Effective Methodology for Principled Integration of Machine Learning and Reasoning (2019, background): “For explainability purposes, another approach couples a deep network with sequence models to extract natural language knowledge [22].”
Explanation in Human-AI Systems: A Literature Meta-Review, Synopsis of Key Ideas and Publications, and Bibliography for Explainable AI (2019, background): “Lingual descriptions of complex systems remain a topic of interest (see Hendricks et al., 2016).”
Learning to Explain With Complemental Examples (2018, background): “As a different stream, some works trained the generative model that outputs explanations with natural language [12, 21] in a post-hoc manner.”
Explainable AI: The New 42? (2018, neutral): “A joint classification and explanation model [7].”
Stacking with Auxiliary Features for Visual Question Answering (2018, background): “Recently, there has been work on analyzing regions of an image that deep-learning models focus on when making decisions (Goyal et al., 2016; Hendricks et al., 2016; Park et al., 2016).”
Deep k-Nearest Neighbors: Towards Confident, Interpretable and Robust Deep Learning (2018, methodology): “follow-ups [58], [59], [60], this technique was also applied to visualize relationships learned between words by the word2vec language model [61].”
Multimodal Explanations: Justifying Decisions and Pointing to the Evidence (2018, methodology): “More recently, [17] developed a deep network to generate natural language justifications of a fine-grained classifier.”