Abstract: In various examples, systems and methods are disclosed relating to jointly pruning channels, layers, and/or blocks of neural networks according to target latency constraints. One or more circuits can determine a plurality of importance scores for a plurality of layers of a neural network and can generate a latency cost data structure for the neural network. The one or more circuits can prune the neural network based at least on the plurality of importance scores, the latency cost data structure, and a target latency value.

Abstract: In various examples, a system can perform multimodal selection of data to generate and/or enrich efficient datasets. The system can retrieve clusters of image frames generated according to semantic characteristics, such as semantic embeddings, of the image frames. The system can selectively filter out image frames from the clusters that are visually similar to other image frames in the clusters, which can reduce the size of the resulting dataset while maintaining target amounts of semantic information in the dataset. The system can selectively add new image frames to the dataset, such as new image frames that have semantic differences from the images of the dataset. The system can update any of various AI models, such as to fine-tune a neural network-based model, suing the dataset.

Abstract: Systems and methods are disclosed for improving natural robustness of sparse neural networks. Pruning a dense neural network may improve inference speed and reduces the memory footprint and energy consumption of the resulting sparse neural network while maintaining a desired level of accuracy. In real-world scenarios in which sparse neural networks deployed in autonomous vehicles perform tasks such as object detection and classification for acquired inputs (images), the neural networks need to be robust to new environments, weather conditions, camera effects, etc. Applying sharpness-aware minimization (SAM) optimization during training of the sparse neural network improves performance for out of distribution (OOD) images compared with using conventional stochastic gradient descent (SGD) optimization. SAM optimizes a neural network to find a flat minimum: a region that both has a small loss value, but that also lies within a region of low loss.

Abstract: Machine learning is a process that learns a neural network model from a given dataset, where the model can then be used to make a prediction about new data. In order to reduce the size, computation, and latency of a neural network model, a compression technique can be employed which includes model sparsification. To avoid the negative consequences of pruning a fully pretrained neural network model and on the other hand of training a sparse model in the first place without any recovery option, the present disclosure provides a dynamic neural network model sparsification process which allows for recovery of previously pruned parts to improve the quality of the sparse neural network model.

Abstract: A technique for dynamically configuring and executing an augmented neural network in real-time according to performance constraints also maintains the legacy neural network execution path. A neural network model that has been trained for a task is augmented with low-compute “shallow” phases paired with each legacy phase and the legacy phases of the neural network model are held constant (e.g., unchanged) while the shallow phases are trained. During inference, one or more of the shallow phases can be selectively executed in place of the corresponding legacy phase. Compared with the legacy phases, the shallow phases are typically less accurate, but have reduced latency and consume less power. Therefore, processing using one or more of the shallow phases in place of one or more of the legacy phases enables the augmented neural network to dynamically adapt to changes in the execution environment (e.g., processing load or performance requirement).

Abstract: Apparatuses, systems, and techniques are presented to simplify neural networks. In at least one embodiment, one or more portions of one or more neural networks are cause to be removed based, at least in part, on one or more performance metrics of the one or more neural networks.

Abstract: Apparatuses, systems, and techniques to remove one or more nodes of a neural network. In at least one embodiment, one or more nodes of a neural network are removed, based on, for example, whether the one or more nodes are likely to affect performance of the neural network.

Abstract: Apparatuses, systems, and techniques for calculating data compression parameters using codebook entry values. In at least one embodiment, one or more circuits is to calculate one or more data compression parameters based, at least in part, on at least on one or more values of the data to be compressed in relation to at least two codebook entry values.