Detecting Scene-Plausible Perceptible Backdoors in Trained DNNs without Access to the Training Set

Backdoor data poisoning attacks add mislabeled examples to the training set, with an embedded backdoor pattern, so that the classifier learns to classify to a target class whenever the backdoor pattern is present in a test sample. Here, we address posttraining detection of scene-plausible perceptible backdoors, a type of backdoor attack that can be relatively easily fashioned, particularly against DNN image classifiers. A posttraining defender does not have access to the potentially poisoned training set, only to the trained classifier, as well as some unpoisoned examples that need not be training samples. Without the poisoned training set, the only information about a backdoor pattern is encoded in the DNN's trained weights. This detection scenario is of great import considering legacy and proprietary systems, cell phone apps, as well as training outsourcing, where the user of the classifier will not have access to the entire training set. We identify two important properties of scene-plausible perceptible backdoor patterns, spatial invariance and robustness, based on which we propose a novel detector using the maximum achievable misclassification fraction (MAMF) statistic. We detect whether the trained DNN has been backdoor-attacked and infer the source and target classes. Our detector outperforms existing detectors and, coupled with an imperceptible backdoor detector, helps achieve posttraining detection of most evasive backdoors of interest.

Deep-seismic-prior-based reconstruction of seismic data using convolutional neural networks

The reconstruction of seismic data with missing traces has been a long-standing issue in seismic data processing; deep learning methods have attracted significant attention in seismic data reconstruction. One barrier associated with these deep-learning based reconstruction methods is the need for large training datasets, which are difficult to acquire owing to physical or financial constraints in practice. A novel method for the recovery of incomplete seismic data without the need of training datasets was developed. Seismic prior is implicitly captured based on the particular CNN structure choice, referred to as the “deep-seismic-prior-based”. The learned network weights are the parameters that represent seismic data, and as the convolutional filter weights are shared for spatial invariance, the CNN structure can function as a regularizer to guide the network learning. The reconstruction is realized during the iterative process by minimizing the mean square error (MSE) between the network output and the original corrupted seismic data. Our method could handle both irregular and regular seismic data, and testing its performance using both synthetic and field data showed it was more advantageous compared with the singular spectrum analysis (SSA) and de-aliased Cadzow methods employed in the reconstruction of irregular and regular data, respectively. The experimental results showed that the proposed method provided better reconstruction performance than the SSA and Cadzow methods.

Exploiting Spatial Invariance for Scalable Unsupervised Object Tracking

The ability to detect and track objects in the visual world is a crucial skill for any intelligent agent, as it is a necessary precursor to any object-level reasoning process. Moreover, it is important that agents learn to track objects without supervision (i.e. without access to annotated training videos) since this will allow agents to begin operating in new environments with minimal human assistance. The task of learning to discover and track objects in videos, which we call unsupervised object tracking, has grown in prominence in recent years; however, most architectures that address it still struggle to deal with large scenes containing many objects. In the current work, we propose an architecture that scales well to the large-scene, many-object setting by employing spatially invariant computations (convolutions and spatial attention) and representations (a spatially local object specification scheme). In a series of experiments, we demonstrate a number of attractive features of our architecture; most notably, that it outperforms competing methods at tracking objects in cluttered scenes with many objects, and that it can generalize well to videos that are larger and/or contain more objects than videos encountered during training.

Learning Spatial Invariance with the Trace Rule in Nonuniform Distributions

Convolutional models of object recognition achieve invariance to spatial transformations largely because of the use of a suitably defined pooling operator. This operator typically takes the form of a max or average function defined across units tuned to the same feature. As a model of the brain's ventral pathway, where computations are carried out by weighted synaptic connections, such pooling can lead to spatial invariance only if the weights that connect similarly tuned units to a given pooling unit are of approximately equal strengths. How identical weights can be learned in the face of nonuniformly distributed data remains unclear. In this letter, we show how various versions of the trace learning rule can help solve this problem. This allows us in turn to explain previously published results and make recommendations as to the optimal rule for invariance learning.