Last Week in AI
Every week, my team at Invector Labs publishes a newsletter to track the most recent developments in AI research and technology. You can find this week’s issue below. You can sign up for it below. Please do so, our guys worked really hard on this:
From the Editor: Drawing Inspiration from the Human Brain
Do artificial intelligence(AI) systems resemble the cognitive patterns of the human brain?
After decades of research in AI, the human brain still remains as a major inspiration for the implementation of new intelligent models. The entire connectionism school is based on creating models based on neural networks inspired by well-known mechanisms in the brain. The deep learning renaissance of the last few years have certainly been the highest point of the connectionist school. However, with the little neuroscience knows today about the human brain, how can we possibly recreate its architecture in AI systems?
When we explore deep learning systems in details, we can clearly see that their similarity with the human brain is more philosophical than technical. Currently, deep learning models have little to do with even the well-known mechanisms of the human brain. However, as both neuroscience and AI evolve, we should expect new neural network architectures to be modeled using brain-like patterns. In a recent example, IBM researchers published a paper proposing a method inspire by the synaptic mechanisms of neuros to optimize the architecture of a neural network. We should expect to see a proliferation of similar efforts in the near future.
Now let’s take a look at the core developments in AI research and technology this week:
AI researchers from IBM published a paper proposing a model that optimized neural networks using inspiration from synaptic mechanisms in the brain.
Researchers from ElementAI published a detailed blog post about modern anomaly detection techniques.
Google AI researchers published a paper proposing a new model for improving the usability of mobile user experiences.
MIT Researchers proposed a new image analysis model that reasons about the contents of a picture.
Cool Tech Releases
Facebook open sourced BigGraph, a framework for graph embeddings in PyTorch.
LinkedIn open sourced Avro2TF, a framework for feature transformation engine for TensorFlow.
Google unveiled a new deep learning model for capturing special moments with Google Photos.
AI in the Real World
Apple just hired Google’s AI researcher Ian Goodfellow, known for the creation of generative adversarial neural networks.
The US Department of Energy(DOE) announced $20 million to develop new machine learning models and tools.
The Guardian published a comprehensive analysis of the risk of AI achieving super-intelligence levels.