2020 Bower Award and Prize for Achievement in Science

The Franklin Institute seeks nominations for the 2020 Bower Award and Prize for Achievement in Science of individuals who have made significant contributions to the development of neural networks for machine learning—hardware or software systems designed as networks of artificial neurons that can be given raw data and trained to automatically discover abstract features that are relevant to detection, classification, or translation, resulting in desired outputs. Inputs could include text (machine translation), audio (speech recognition), or imagery (face recognition, scene understanding, photo sorting, image synthesis). Outputs could include categorical labels, structured outputs, or actuator commands. Such artificial neural networks have yielded effective approaches to solving a wide spectrum of challenging practical problems resistant to solution using earlier algorithmic machine learning techniques.

Nominations are encouraged in, but not limited to, the following subtopics:

• Network architectures: layered, recurrent, self-organizing, convolutional, long short-term memory, auto encoders, deep networks
• Training or learning methods: cooperative, competitive, back propagation of errors, supervised, unsupervised, reinforcement, adaptive resonance, generative adversarial
• Network implementation: analog, digital, stochastic, energy-efficient, hardware-accelerated
• Broad applications: speech recognition, language translation, face recognition, scene understanding, photo sorting, reading print or handwriting, medical image diagnosis, autonomous vehicles, drug delivery design, question answering, web searches, game playing