Inferring Motor Programs from Images of Handwritten Digits

Neural Information Processing Systems

We describe a generative model for handwritten digits that uses two pairs of opposing springs whose stiffnesses are controlled by a motor program. We show how neural networks can be trained to infer the motor programs required to accurately reconstruct the MNIST digits. The inferred motor programs can be used directly for digit classification, but they can also be used in other ways. By adding noise to the motor program inferred from an MNIST image we can generate a large set of very different images of the same class, thus enlarging the training set available to other methods. We can also use the motor programs as additional, highly informative outputs which reduce overfitting when training a feed-forward classifier.


Shape and Material from Sound

Neural Information Processing Systems

Hearing an object falling onto the ground, humans can recover rich information including its rough shape, material, and falling height. In this paper, we build machines to approximate such competency. We first mimic human knowledge of the physical world by building an efficient, physics-based simulation engine. Then, we present an analysis-by-synthesis approach to infer properties of the falling object. We further accelerate the process by learning a mapping from a sound wave to object properties, and using the predicted values to initialize the inference. This mapping can be viewed as an approximation of human commonsense learned from past experience. Our model performs well on both synthetic audio clips and real recordings without requiring any annotated data. We conduct behavior studies to compare human responses with ours on estimating object shape, material, and falling height from sound. Our model achieves near-human performance.


Mapping the Brain to Build Better Machines Quanta Magazine

#artificialintelligence

Take a three year-old to the zoo, and she intuitively knows that the long-necked creature nibbling leaves is the same thing as the giraffe in her picture book. That superficially easy feat is in reality quite sophisticated. The cartoon drawing is a frozen silhouette of simple lines, while the living animal is awash in color, texture, movement and light. It can contort into different shapes and looks different from every angle. Humans excel at this kind of task.


Mapping the Brain to Build Better Machines Quanta Magazine

#artificialintelligence

Take a three year-old to the zoo, and she intuitively knows that the long-necked creature nibbling leaves is the same thing as the giraffe in her picture book. That superficially easy feat is in reality quite sophisticated. The cartoon drawing is a frozen silhouette of simple lines, while the living animal is awash in color, texture, movement and light. It can contort into different shapes and looks different from every angle. Humans excel at this kind of task.


Attentional Neural Network: Feature Selection Using Cognitive Feedback

Neural Information Processing Systems

Attentional Neural Network is a new framework that integrates top-down cognitive bias and bottom-up feature extraction in one coherent architecture. The top-down influence is especially effective when dealing with high noise or difficult segmentation problems. Our system is modular and extensible. It is also easy to train and cheap to run, and yet can accommodate complex behaviors. We obtain classification accuracy better than or competitive with state of art results on the MNIST variation dataset, and successfully disentangle overlaid digits with high success rates. We view such a general purpose framework as an essential foundation for a larger system emulating the cognitive abilities of the whole brain.