The utility of Neural Networks in Artificial Intelligence has been the domain of wild predictions and conjecture for many decades. After an initially overabundance of optimism in late 80s and 90s they fell out of favor as the computing platforms available failed to live up to the expected performance. Forgotten largely except for a few die-hard researchers, a specific variant of Artificial Neural Networks known as Deep Learning is making a comeback in a big way.

Thanks to new computing paradigms, abundance of Big data, better hardware such as powerful GPUs as well as proof of concept cases proving their effectiveness, Deep Learning is transforming what is possible in artificial intelligence in image and speech recognition, machine translation and machine learning.

At IC, in concert with big data platforms and the latest technology from academia, we are applying Deep Learning to business problems attempting to solve strategic problems that have been a black art or had never been attempted due to their prohibitive complexity for the human mind alone.

IC Predictive Logistics applies Deep Learning Artificial Neural Network technology to business problems with the goal of producing "super human" result for your strategic business problems.

Ingesting large volumes of labeled and unlabeled data, the latest technics in this fast moving filed, widely available open source software and cloud infrastructure as a service platforms, we are ushering in the next era of business strategy formulation and business optimization problems. Unlike in the past where predictive analytics was largely dependent on what is known as feature engineering where researchers tried to use their domain expertise to come up with features (data attributes) that they deem important to the predictions, the latest technics use unsupervised learning and q-learning to allow the Neural Net to self-train on data to find the features that are predictive (important) to the outcome and optimize the model for a given strategic objective.

With deep reinforcement learning and memory based neural networks, we are talking the next logical step in this evolution by allowing the data influence and teach the network's learning strategy in such a way that the system is allowed to invent ways of maximizing strategic long term outcome for a problem by gradually adopting super-human strategies unseen by the human mind. No amount of analytics or dashboards or "insights" will ever be able to match the sheer computational efficiency of such learning models. This is the essence of Predictive Logistics.

Leveraging our long history with Extraction Transformation and Loading (ETL) platform development, we will apply these models to business problems in the presence of structured and unstructured data and let your custom learning system on it's own maximize strategic outcomes by learning the strategy from the data points via long-term iterative learning and policy memory.

We believe that "Insights from Data" is yesterday's solution, dashboards are just pretty pictures with very little strategic value and that analytics are only as useful as the producer's mind and the consumer's eye.

Predictive Logistics from International Computing is Strategy from Data