SignificanceThe ability to complete patterns and interpret partial information is a central property of intelligence. Deep convolutional network architectures have proved successful in labeling whole objects in images and capturing the initial 150 ms of processing along the ventral visual cortex. This study shows that human object recognition abilities remain robust when only small amounts of information are available due to heavy occlusion, but the performance of bottom-up computational models is impaired under limited visibility. The results provide combined behavioral, neurophysiological, and modeling insights showing how recurrent computations may help the brain solve the fundamental challenge of pattern completion.

Summary Natural vision often involves recognizing objects from partial information. Recognition of objects from parts presents a significant challenge for theories of vision because it requires spatial integration and extrapolation from prior knowledge. Here we recorded intracranial field potentials of 113 visually selective electrodes from epilepsy patients in response to whole and partial objects. Responses along the ventral visual stream, particularly the Inferior Occipital and Fusiform Gyri, remained selective despite showing only 9–25% of the object areas. However, these visually selective signals emerged ~100 ms later for partial versus whole objects. These processing delays were particularly pronounced in higher visual areas within the ventral stream. This latency difference persisted when controlling for changes in contrast, signal amplitude, and the strength of selectivity. These results argue against a purely feed-forward explanation of recognition from partial information, and provide spatiotemporal constraints on theories of object recognition that involve recurrent processing.

Machine-learning (ML) hardware and software system demand is burgeoning. Driven by ML applications, the number of different ML inference systems has exploded. Over 100 organizations are building ML inference chips, and the systems that incorporate existing models span at least three orders of magnitude in power consumption and four orders of magnitude in performance; they range from embedded devices to data-center solutions. Fueling the hardware are a dozen or more software frameworks and libraries. The myriad combinations of ML hardware and ML software make assessing ML-system performance in an architecture-neutral, representative, and reproducible manner challenging. There is a clear need for industry-wide standard ML benchmarking and evaluation criteria. MLPerf Inference answers that call. Driven by more than 30 organizations as well as more than 200 ML engineers and practitioners, MLPerf implements a set of rules and practices to ensure comparability across systems with wildly differing architectures. In this paper, we present the method and design principles of the initial MLPerf Inference release. The first call for submissions garnered more than 600 inference-performance measurements from 14 organizations, representing over 30 systems that show a range of capabilities.