Supervised learning in artificial neural networks typically relies on backpropagation, where the weights are updated based on the error-function gradients and sequentially propagated from the output layer to the input layer. Although this approach has proven effective in a wide domain of applications, it lacks biological plausibility in many regards, including the weight symmetry problem, the dependence of learning on non-local signals, the freezing of neural activity during error propagation, and the update locking problem. Alternative training schemes - such as sign symmetry, feedback alignment, and direct feedback alignment - have been introduced, but invariably rely on a backward pass that hinders the possibility of solving all the issues simultaneously. Here, we propose to replace the backward pass with a second forward pass in which the input signal is modulated based on the error of the network. We show that this novel learning rule comprehensively addresses all the above-mentioned issues and can be applied to both fully connected and convolutional models. We test this learning rule on MNIST, CIFAR-10, and CIFAR-100. These results help incorporate biological principles into machine learning.
In recent years, multi-modal transformers have shown significant progress in Vision-Language tasks, such as Visual Question Answering (VQA), outperforming previous architectures by a considerable margin. This improvement in VQA is often attributed to the rich interactions between vision and language streams. In this work, we investigate the efficacy of co-attention transformer layers in helping the network focus on relevant regions while answering the question. We generate visual attention maps using the question-conditioned image attention scores in these co-attention layers. We evaluate the effect of the following critical components on visual attention of a state-of-the-art VQA model: (i) number of object region proposals, (ii) question part of speech (POS) tags, (iii) question semantics, (iv) number of co-attention layers, and (v) answer accuracy. We compare the neural network attention maps against human attention maps both qualitatively and quantitatively. Our findings indicate that co-attention transformer modules are crucial in attending to relevant regions of the image given a question. Importantly, we observe that the semantic meaning of the question is not what drives visual attention, but specific keywords in the question do. Our work sheds light on the function and interpretation of co-attention transformer layers, highlights gaps in current networks, and can guide the development of future VQA models and networks that simultaneously process visual and language streams.
It is well understood that modern deep networks are vulnerable to adversarial attacks. However, conventional methods fail to produce adversarial perturbations that are intelligible to humans, and they pose limited threats in the physical world. To study feature-class associations in networks and better understand the real-world threats they face, we develop feature-level adversarial perturbations using deep image generators and a novel optimization objective. We term these feature-fool attacks. We show that they are versatile and use them to generate targeted feature-level attacks at the ImageNet scale that are simultaneously interpretable, universal to any source image, and physically-realizable. These attacks can also reveal spurious, semantically-describable feature/class associations, and we use them to guide the design of "copy/paste" adversaries in which one natural image is pasted into another to cause a targeted misclassification.
Visual search is a ubiquitous and often challenging daily task, exemplified by looking for the car keys at home or a friend in a crowd. An intriguing property of some classical search tasks is an asymmetry such that finding a target A among distractors B can be easier than finding B among A. To elucidate the mechanisms responsible for asymmetry in visual search, we propose a computational model that takes a target and a search image as inputs and produces a sequence of eye movements until the target is found. The model integrates eccentricity-dependent visual recognition with target-dependent top-down cues. We compared the model against human behavior in six paradigmatic search tasks that show asymmetry in humans. Without prior exposure to the stimuli or task-specific training, the model provides a plausible mechanism for search asymmetry. We hypothesized that the polarity of search asymmetry arises from experience with the natural environment. We tested this hypothesis by training the model on an augmented version of ImageNet where the biases of natural images were either removed or reversed. The polarity of search asymmetry disappeared or was altered depending on the training protocol. This study highlights how classical perceptual properties can emerge in neural network models, without the need for task-specific training, but rather as a consequence of the statistical properties of the developmental diet fed to the model. All source code and stimuli are publicly available https://github.com/kreimanlab/VisualSearchAsymmetry
In human vision objects and their parts can be visually recognized from purely spatial or purely temporal information but the mechanisms integrating space and time are poorly understood. Here we show that human visual recognition of objects and actions can be achieved by efficiently combining spatial and motion cues in configurations where each source on its own is insufficient for recognition. This analysis is obtained by identifying minimal videos: these are short and tiny video clips in which objects, parts, and actions can be reliably recognized, but any reduction in either space or time makes them unrecognizable. State-of-the-art deep networks for dynamic visual recognition cannot replicate human behavior in these configurations. This gap between humans and machines points to critical mechanisms in human dynamic vision that are lacking in current models.
Stream learning refers to the ability to acquire and transfer knowledge across a continuous stream of data without forgetting and without repeated passes over the data. A common way to avoid catastrophic forgetting is to intersperse new examples with replays of old examples stored as image pixels or reproduced by generative models. Here, we considered stream learning in image classification tasks and proposed a novel hypotheses-driven Augmented Memory Network, which efficiently consolidates previous knowledge with a limited number of hypotheses in the augmented memory and replays relevant hypotheses to avoid catastrophic forgetting. The advantages of hypothesis-driven replay over image pixel replay and generative replay are two-fold. First, hypothesis-based knowledge consolidation avoids redundant information in the image pixel space and makes memory usage more efficient. Second, hypotheses in the augmented memory can be re-used for learning new tasks, improving generalization and transfer learning ability. We evaluated our method on three stream learning object recognition datasets. Our method performs comparably well or better than SOTA methods, while offering more efficient memory usage. All source code and data are publicly available https://github.com/kreimanlab/AugMem.
Context is of fundamental importance to both human and machine vision -- an object in the air is more likely to be an airplane, than a pig. The rich notion of context incorporates several aspects including physics rules, statistical co-occurrences, and relative object sizes, among others. While previous works have crowd-sourced out-of-context photographs from the web to study scene context, controlling the nature and extent of contextual violations has been an extremely daunting task. Here we introduce a diverse, synthetic Out-of-Context Dataset (OCD) with fine-grained control over scene context. By leveraging a 3D simulation engine, we systematically control the gravity, object co-occurrences and relative sizes across 36 object categories in a virtual household environment. We then conduct a series of experiments to gain insights into the impact of contextual cues on both human and machine vision using OCD. First, we conduct psycho-physics experiments to establish a human benchmark for out-of-context recognition, and then compare it with state-of-the-art computer vision models to quantify the gap between the two. Finally, we propose a context-aware recognition transformer model, fusing object and contextual information via multi-head attention. Our model captures useful information for contextual reasoning, enabling human-level performance and significantly better robustness in out-of-context conditions compared to baseline models across OCD and other existing out-of-context natural image datasets. All source code and data are publicly available https://github.com/kreimanlab/WhenPigsFlyContext.
Saccadic eye movements allow animals to bring different parts of an image into high-resolution. During free viewing, inhibition of return incentivizes exploration by discouraging previously visited locations. Despite this inhibition, here we show that subjects make frequent return fixations. We systematically studied a total of 44,328 return fixations out of 217,440 fixations across different tasks, in monkeys and humans, and in static images or egocentric videos. The ubiquitous return fixations were consistent across subjects, tended to occur within short offsets, and were characterized by longer duration than non-return fixations. The locations of return fixations corresponded to image areas of higher saliency and higher similarity to the sought target during visual search tasks. We propose a biologically-inspired computational model that capitalizes on a deep convolutional neural network for object recognition to predict a sequence of fixations. Given an input image, the model computes four maps that constrain the location of the next saccade: a saliency map, a target similarity map, a saccade size map, and a memory map. The model exhibits frequent return fixations and approximates the properties of return fixations across tasks and species. The model provides initial steps towards capturing the trade-off between exploitation of informative image locations combined with exploration of novel image locations during scene viewing.
Deep artificial neural networks have been proposed as a model of primate vision. However, these networks are vulnerable to adversarial attacks, whereby introducing minimal noise can fool networks into misclassifying images. Primate vision is thought to be robust to such adversarial images. We evaluated this assumption by designing adversarial images to fool primate vision. To do so, we first trained a model to predict responses of face-selective neurons in macaque inferior temporal cortex. Next, we modified images, such as human faces, to match their model-predicted neuronal responses to a target category, such as monkey faces. These adversarial images elicited neuronal responses similar to the target category. Remarkably, the same images fooled monkeys and humans at the behavioral level. These results challenge fundamental assumptions about the similarity between computer and primate vision and show that a model of neuronal activity can selectively direct primate visual behavior.
Can we infer intentions from a person's actions? As an example problem, here we consider how to decipher what a person is searching for by decoding their eye movement behavior. We conducted two psychophysics experiments where we monitored eye movements while subjects searched for a target object. We defined the fixations falling on \textit{non-target} objects as "error fixations". Using those error fixations, we developed a model (InferNet) to infer what the target was. InferNet uses a pre-trained convolutional neural network to extract features from the error fixations and computes a similarity map between the error fixations and all locations across the search image. The model consolidates the similarity maps across layers and integrates these maps across all error fixations. InferNet successfully identifies the subject's goal and outperforms competitive null models, even without any object-specific training on the inference task.