Neural Information Processing Systems http://nips.cc/

Accurate short-term streamflow and flood forecasting are critical for mitigating river flood impacts, especially given the increasing climate variability. Machine learning-based streamflow forecasting relies on large streamflow datasets derived from rating curves. Uncertainties in rating curve modeling could introduce errors to the streamflow data and affect the forecasting accuracy. This study proposes a streamflow forecasting method that addresses these data errors, enhancing the accuracy of river flood forecasting and flood modeling, thereby reducing flood-related risk. A convolutional recurrent neural network is used to capture spatiotemporal patterns, coupled with residual error learning and forecasting. The neural network outperforms commonly used forecasting models over 1-6 hours of forecasting horizons, and the residual error learners can further correct the residual errors. This provides a more reliable tool for river flood forecasting and climate adaptation in this critical 1-6 hour time window for flood risk mitigation efforts.

We study how well one can recover sparse principal components of a data matrix using a sketch formed from a few of its elements. We show that for a wide class of optimization problems, if the sketch is close (in the spectral norm) to the original data matrix, then one can recover a near optimal solution to the optimization problem by using the sketch.

The Vanderburgh County Coroner's office says Austin Smith died Friday. He was shot on Aug. 31, 2017. Twenty-two-year-old Travis Phelps is accused of firing several shots into Smith's car, causing him to crash.

Pienso, a machine learning platform for non-programmers, has closed a $2.1 million seed round. Led by Eniac Ventures, with participation from SoftTech VC, Indicator Ventures and E14 Fund, Pienso is focused on democratizing machine learning for domain experts who are non-programmers with no technical or data scientist experience. The funding allows the company to scale operations. "Investment by large enterprises in machine learning is rapidly accelerating as corporations spin up massive data lakes to garner insights into their business. However, it is costly, challenging to integrate and before now required data scientists on staff," said Vic Singh, Indian American general partner at Eniac.

This paper is part of a larger project to determine how to build agent-based cognitive models capable of initial associative intelligence. Our method here is to take McClelland’s 1981 “Jets and Sharks” dataset and rebuild it using a nonlinear dynamic system with an eye toward determining which parameters are necessary to govern the interactivity of agents in a multi-agent cognitive system. A few number of parameters are suggested concerning diffusion and infusion values, which are basically elementary forms of information entropy, and multi-dimensional overlap from properties to objects and then from objects back to the properties that define them. While no agent-based model is presented, the success of the dynamic systems that are presented here suggest strong starting points for further research in building cognitive complex adaptive systems.