Models, code, and papers for "Antonio Torralba":

Notes on image annotation

Oct 12, 2012
Adela Barriuso, Antonio Torralba

We are under the illusion that seeing is effortless, but frequently the visual system is lazy and makes us believe that we understand something when in fact we don't. Labeling a picture forces us to become aware of the difficulties underlying scene understanding. Suddenly, the act of seeing is not effortless anymore. We have to make an effort in order to understand parts of the picture that we neglected at first glance. In this report, an expert image annotator relates her experience on segmenting and labeling tens of thousands of images. During this process, the notes she took try to highlight the difficulties encountered, the solutions adopted, and the decisions made in order to get a consistent set of annotations. Those annotations constitute the SUN database.

* 15 pages 

  Click for Model/Code and Paper
See, Hear, and Read: Deep Aligned Representations

Jun 03, 2017
Yusuf Aytar, Carl Vondrick, Antonio Torralba

We capitalize on large amounts of readily-available, synchronous data to learn a deep discriminative representations shared across three major natural modalities: vision, sound and language. By leveraging over a year of sound from video and millions of sentences paired with images, we jointly train a deep convolutional network for aligned representation learning. Our experiments suggest that this representation is useful for several tasks, such as cross-modal retrieval or transferring classifiers between modalities. Moreover, although our network is only trained with image+text and image+sound pairs, it can transfer between text and sound as well, a transfer the network never observed during training. Visualizations of our representation reveal many hidden units which automatically emerge to detect concepts, independent of the modality.


  Click for Model/Code and Paper
Who is Mistaken?

Mar 31, 2017
Benjamin Eysenbach, Carl Vondrick, Antonio Torralba

Recognizing when people have false beliefs is crucial for understanding their actions. We introduce the novel problem of identifying when people in abstract scenes have incorrect beliefs. We present a dataset of scenes, each visually depicting an 8-frame story in which a character has a mistaken belief. We then create a representation of characters' beliefs for two tasks in human action understanding: predicting who is mistaken, and when they are mistaken. Experiments suggest that our method for identifying mistaken characters performs better on these tasks than simple baselines. Diagnostics on our model suggest it learns important cues for recognizing mistaken beliefs, such as gaze. We believe models of people's beliefs will have many applications in action understanding, robotics, and healthcare.

* See project website at: http://people.csail.mit.edu/bce/mistaken/ . (Edit: fixed typos and references) 

  Click for Model/Code and Paper
Anticipating Visual Representations from Unlabeled Video

Nov 30, 2016
Carl Vondrick, Hamed Pirsiavash, Antonio Torralba

Anticipating actions and objects before they start or appear is a difficult problem in computer vision with several real-world applications. This task is challenging partly because it requires leveraging extensive knowledge of the world that is difficult to write down. We believe that a promising resource for efficiently learning this knowledge is through readily available unlabeled video. We present a framework that capitalizes on temporal structure in unlabeled video to learn to anticipate human actions and objects. The key idea behind our approach is that we can train deep networks to predict the visual representation of images in the future. Visual representations are a promising prediction target because they encode images at a higher semantic level than pixels yet are automatic to compute. We then apply recognition algorithms on our predicted representation to anticipate objects and actions. We experimentally validate this idea on two datasets, anticipating actions one second in the future and objects five seconds in the future.

* CVPR 2016 

  Click for Model/Code and Paper
SoundNet: Learning Sound Representations from Unlabeled Video

Oct 27, 2016
Yusuf Aytar, Carl Vondrick, Antonio Torralba

We learn rich natural sound representations by capitalizing on large amounts of unlabeled sound data collected in the wild. We leverage the natural synchronization between vision and sound to learn an acoustic representation using two-million unlabeled videos. Unlabeled video has the advantage that it can be economically acquired at massive scales, yet contains useful signals about natural sound. We propose a student-teacher training procedure which transfers discriminative visual knowledge from well established visual recognition models into the sound modality using unlabeled video as a bridge. Our sound representation yields significant performance improvements over the state-of-the-art results on standard benchmarks for acoustic scene/object classification. Visualizations suggest some high-level semantics automatically emerge in the sound network, even though it is trained without ground truth labels.

* NIPS 2016 

  Click for Model/Code and Paper
Generating Videos with Scene Dynamics

Oct 26, 2016
Carl Vondrick, Hamed Pirsiavash, Antonio Torralba

We capitalize on large amounts of unlabeled video in order to learn a model of scene dynamics for both video recognition tasks (e.g. action classification) and video generation tasks (e.g. future prediction). We propose a generative adversarial network for video with a spatio-temporal convolutional architecture that untangles the scene's foreground from the background. Experiments suggest this model can generate tiny videos up to a second at full frame rate better than simple baselines, and we show its utility at predicting plausible futures of static images. Moreover, experiments and visualizations show the model internally learns useful features for recognizing actions with minimal supervision, suggesting scene dynamics are a promising signal for representation learning. We believe generative video models can impact many applications in video understanding and simulation.

* NIPS 2016. See more at http://web.mit.edu/vondrick/tinyvideo/ 

  Click for Model/Code and Paper
The Role of Embedding Complexity in Domain-invariant Representations

Oct 13, 2019
Ching-Yao Chuang, Antonio Torralba, Stefanie Jegelka

Unsupervised domain adaptation aims to generalize the hypothesis trained in a source domain to an unlabeled target domain. One popular approach to this problem is to learn domain-invariant embeddings for both domains. In this work, we study, theoretically and empirically, the effect of the embedding complexity on generalization to the target domain. In particular, this complexity affects an upper bound on the target risk; this is reflected in experiments, too. Next, we specify our theoretical framework to multilayer neural networks. As a result, we develop a strategy that mitigates sensitivity to the embedding complexity, and empirically achieves performance on par with or better than the best layer-dependent complexity tradeoff.


  Click for Model/Code and Paper
Temporal Relational Reasoning in Videos

Jul 25, 2018
Bolei Zhou, Alex Andonian, Aude Oliva, Antonio Torralba

Temporal relational reasoning, the ability to link meaningful transformations of objects or entities over time, is a fundamental property of intelligent species. In this paper, we introduce an effective and interpretable network module, the Temporal Relation Network (TRN), designed to learn and reason about temporal dependencies between video frames at multiple time scales. We evaluate TRN-equipped networks on activity recognition tasks using three recent video datasets - Something-Something, Jester, and Charades - which fundamentally depend on temporal relational reasoning. Our results demonstrate that the proposed TRN gives convolutional neural networks a remarkable capacity to discover temporal relations in videos. Through only sparsely sampled video frames, TRN-equipped networks can accurately predict human-object interactions in the Something-Something dataset and identify various human gestures on the Jester dataset with very competitive performance. TRN-equipped networks also outperform two-stream networks and 3D convolution networks in recognizing daily activities in the Charades dataset. Further analyses show that the models learn intuitive and interpretable visual common sense knowledge in videos.

* camera-ready version for ECCV'18 

  Click for Model/Code and Paper
Interpreting Deep Visual Representations via Network Dissection

Jun 26, 2018
Bolei Zhou, David Bau, Aude Oliva, Antonio Torralba

The success of recent deep convolutional neural networks (CNNs) depends on learning hidden representations that can summarize the important factors of variation behind the data. However, CNNs often criticized as being black boxes that lack interpretability, since they have millions of unexplained model parameters. In this work, we describe Network Dissection, a method that interprets networks by providing labels for the units of their deep visual representations. The proposed method quantifies the interpretability of CNN representations by evaluating the alignment between individual hidden units and a set of visual semantic concepts. By identifying the best alignments, units are given human interpretable labels across a range of objects, parts, scenes, textures, materials, and colors. The method reveals that deep representations are more transparent and interpretable than expected: we find that representations are significantly more interpretable than they would be under a random equivalently powerful basis. We apply the method to interpret and compare the latent representations of various network architectures trained to solve different supervised and self-supervised training tasks. We then examine factors affecting the network interpretability such as the number of the training iterations, regularizations, different initializations, and the network depth and width. Finally we show that the interpreted units can be used to provide explicit explanations of a prediction given by a CNN for an image. Our results highlight that interpretability is an important property of deep neural networks that provides new insights into their hierarchical structure.

* *B. Zhou and D. Bau contributed equally to this work. 15 pages, 27 figures 

  Click for Model/Code and Paper
Revisiting the Importance of Individual Units in CNNs via Ablation

Jun 07, 2018
Bolei Zhou, Yiyou Sun, David Bau, Antonio Torralba

We revisit the importance of the individual units in Convolutional Neural Networks (CNNs) for visual recognition. By conducting unit ablation experiments on CNNs trained on large scale image datasets, we demonstrate that, though ablating any individual unit does not hurt overall classification accuracy, it does lead to significant damage on the accuracy of specific classes. This result shows that an individual unit is specialized to encode information relevant to a subset of classes. We compute the correlation between the accuracy drop under unit ablation and various attributes of an individual unit such as class selectivity and weight L1 norm. We confirm that unit attributes such as class selectivity are a poor predictor for impact on overall accuracy as found previously in recent work \cite{morcos2018importance}. However, our results show that class selectivity along with other attributes are good predictors of the importance of one unit to individual classes. We evaluate the impact of random rotation, batch normalization, and dropout to the importance of units to specific classes. Our results show that units with high selectivity play an important role in network classification power at the individual class level. Understanding and interpreting the behavior of these units is necessary and meaningful.


  Click for Model/Code and Paper
Following Gaze Across Views

Dec 09, 2016
Adrià Recasens, Carl Vondrick, Aditya Khosla, Antonio Torralba

Following the gaze of people inside videos is an important signal for understanding people and their actions. In this paper, we present an approach for following gaze across views by predicting where a particular person is looking throughout a scene. We collect VideoGaze, a new dataset which we use as a benchmark to both train and evaluate models. Given one view with a person in it and a second view of the scene, our model estimates a density for gaze location in the second view. A key aspect of our approach is an end-to-end model that solves the following sub-problems: saliency, gaze pose, and geometric relationships between views. Although our model is supervised only with gaze, we show that the model learns to solve these subproblems automatically without supervision. Experiments suggest that our approach follows gaze better than standard baselines and produces plausible results for everyday situations.

* 9 pages, 8 figures 

  Click for Model/Code and Paper
Predicting Motivations of Actions by Leveraging Text

Nov 30, 2016
Carl Vondrick, Deniz Oktay, Hamed Pirsiavash, Antonio Torralba

Understanding human actions is a key problem in computer vision. However, recognizing actions is only the first step of understanding what a person is doing. In this paper, we introduce the problem of predicting why a person has performed an action in images. This problem has many applications in human activity understanding, such as anticipating or explaining an action. To study this problem, we introduce a new dataset of people performing actions annotated with likely motivations. However, the information in an image alone may not be sufficient to automatically solve this task. Since humans can rely on their lifetime of experiences to infer motivation, we propose to give computer vision systems access to some of these experiences by using recently developed natural language models to mine knowledge stored in massive amounts of text. While we are still far away from fully understanding motivation, our results suggest that transferring knowledge from language into vision can help machines understand why people in images might be performing an action.

* CVPR 2016 

  Click for Model/Code and Paper
Learning visual biases from human imagination

Nov 16, 2015
Carl Vondrick, Hamed Pirsiavash, Aude Oliva, Antonio Torralba

Although the human visual system can recognize many concepts under challenging conditions, it still has some biases. In this paper, we investigate whether we can extract these biases and transfer them into a machine recognition system. We introduce a novel method that, inspired by well-known tools in human psychophysics, estimates the biases that the human visual system might use for recognition, but in computer vision feature spaces. Our experiments are surprising, and suggest that classifiers from the human visual system can be transferred into a machine with some success. Since these classifiers seem to capture favorable biases in the human visual system, we further present an SVM formulation that constrains the orientation of the SVM hyperplane to agree with the bias from human visual system. Our results suggest that transferring this human bias into machines may help object recognition systems generalize across datasets and perform better when very little training data is available.

* To appear at NIPS 2015 

  Click for Model/Code and Paper
Are all training examples equally valuable?

Nov 25, 2013
Agata Lapedriza, Hamed Pirsiavash, Zoya Bylinskii, Antonio Torralba

When learning a new concept, not all training examples may prove equally useful for training: some may have higher or lower training value than others. The goal of this paper is to bring to the attention of the vision community the following considerations: (1) some examples are better than others for training detectors or classifiers, and (2) in the presence of better examples, some examples may negatively impact performance and removing them may be beneficial. In this paper, we propose an approach for measuring the training value of an example, and use it for ranking and greedily sorting examples. We test our methods on different vision tasks, models, datasets and classifiers. Our experiments show that the performance of current state-of-the-art detectors and classifiers can be improved when training on a subset, rather than the whole training set.


  Click for Model/Code and Paper
Inverting and Visualizing Features for Object Detection

May 05, 2013
Carl Vondrick, Aditya Khosla, Tomasz Malisiewicz, Antonio Torralba

We introduce algorithms to visualize feature spaces used by object detectors. The tools in this paper allow a human to put on `HOG goggles' and perceive the visual world as a HOG based object detector sees it. We found that these visualizations allow us to analyze object detection systems in new ways and gain new insight into the detector's failures. For example, when we visualize the features for high scoring false alarms, we discovered that, although they are clearly wrong in image space, they do look deceptively similar to true positives in feature space. This result suggests that many of these false alarms are caused by our choice of feature space, and indicates that creating a better learning algorithm or building bigger datasets is unlikely to correct these errors. By visualizing feature spaces, we can gain a more intuitive understanding of our detection systems.

* This paper is a preprint of our conference paper. We have made it available early in the hopes that others find it useful 

  Click for Model/Code and Paper
Learning Compositional Koopman Operators for Model-Based Control

Oct 18, 2019
Yunzhu Li, Hao He, Jiajun Wu, Dina Katabi, Antonio Torralba

Finding an embedding space for a linear approximation of a nonlinear dynamical system enables efficient system identification and control synthesis. The Koopman operator theory lays the foundation for identifying the nonlinear-to-linear coordinate transformations with data-driven methods. Recently, researchers have proposed to use deep neural networks as a more expressive class of basis functions for calculating the Koopman operators. These approaches, however, assume a fixed dimensional state space; they are therefore not applicable to scenarios with a variable number of objects. In this paper, we propose to learn compositional Koopman operators, using graph neural networks to encode the state into object-centric embeddings and using a block-wise linear transition matrix to regularize the shared structure across objects. The learned dynamics can quickly adapt to new environments of unknown physical parameters and produce control signals to achieve a specified goal. Our experiments on manipulating ropes and controlling soft robots show that the proposed method has better efficiency and generalization ability than existing baselines.

* The first two authors contributed equally to this paper. Project Page: http://koopman.csail.mit.edu/ Video: https://www.youtube.com/watch?v=idFH4K16cfQ&feature=youtu.be 

  Click for Model/Code and Paper
Connecting Touch and Vision via Cross-Modal Prediction

Jun 14, 2019
Yunzhu Li, Jun-Yan Zhu, Russ Tedrake, Antonio Torralba

Humans perceive the world using multi-modal sensory inputs such as vision, audition, and touch. In this work, we investigate the cross-modal connection between vision and touch. The main challenge in this cross-domain modeling task lies in the significant scale discrepancy between the two: while our eyes perceive an entire visual scene at once, humans can only feel a small region of an object at any given moment. To connect vision and touch, we introduce new tasks of synthesizing plausible tactile signals from visual inputs as well as imagining how we interact with objects given tactile data as input. To accomplish our goals, we first equip robots with both visual and tactile sensors and collect a large-scale dataset of corresponding vision and tactile image sequences. To close the scale gap, we present a new conditional adversarial model that incorporates the scale and location information of the touch. Human perceptual studies demonstrate that our model can produce realistic visual images from tactile data and vice versa. Finally, we present both qualitative and quantitative experimental results regarding different system designs, as well as visualizing the learned representations of our model.

* Accepted to CVPR 2019. Project Page: http://visgel.csail.mit.edu/ 

  Click for Model/Code and Paper
The Sound of Motions

Apr 11, 2019
Hang Zhao, Chuang Gan, Wei-Chiu Ma, Antonio Torralba

Sounds originate from object motions and vibrations of surrounding air. Inspired by the fact that humans is capable of interpreting sound sources from how objects move visually, we propose a novel system that explicitly captures such motion cues for the task of sound localization and separation. Our system is composed of an end-to-end learnable model called Deep Dense Trajectory (DDT), and a curriculum learning scheme. It exploits the inherent coherence of audio-visual signals from a large quantities of unlabeled videos. Quantitative and qualitative evaluations show that comparing to previous models that rely on visual appearance cues, our motion based system improves performance in separating musical instrument sounds. Furthermore, it separates sound components from duets of the same category of instruments, a challenging problem that has not been addressed before.


  Click for Model/Code and Paper
Dataset Distillation

Nov 27, 2018
Tongzhou Wang, Jun-Yan Zhu, Antonio Torralba, Alexei A. Efros

Model distillation aims to distill the knowledge of a complex model into a simpler one. In this paper, we consider an alternative formulation called {\em dataset distillation}: we keep the model fixed and instead attempt to distill the knowledge from a large training dataset into a small one. The idea is to {\em synthesize} a small number of data points that do not need to come from the correct data distribution, but will, when given to the learning algorithm as training data, approximate the model trained on the original data. For example, we show that it is possible to compress $60,000$ MNIST training images into just $10$ synthetic {\em distilled images} (one per class) and achieve close to original performance with only a few steps of gradient descent, given a particular fixed network initialization. We evaluate our method in a wide range of initialization settings and with different learning objectives. Experiments on multiple datasets show the advantage of our approach compared to alternative methods in most settings.


  Click for Model/Code and Paper
Learning to Act Properly: Predicting and Explaining Affordances from Images

Jun 15, 2018
Ching-Yao Chuang, Jiaman Li, Antonio Torralba, Sanja Fidler

We address the problem of affordance reasoning in diverse scenes that appear in the real world. Affordances relate the agent's actions to their effects when taken on the surrounding objects. In our work, we take the egocentric view of the scene, and aim to reason about action-object affordances that respect both the physical world as well as the social norms imposed by the society. We also aim to teach artificial agents why some actions should not be taken in certain situations, and what would likely happen if these actions would be taken. We collect a new dataset that builds upon ADE20k, referred to as ADE-Affordance, which contains annotations enabling such rich visual reasoning. We propose a model that exploits Graph Neural Networks to propagate contextual information from the scene in order to perform detailed affordance reasoning about each object. Our model is showcased through various ablation studies, pointing to successes and challenges in this complex task.


  Click for Model/Code and Paper