Lucidity Sciences

What if your data wasn't the problem? While there are lots of things that can derail an ML project, you might not have considered one of the most unavoidable ones: limitations in the underlying ML algorithm. Why it matters, and a way forward. #machinelearning #ml #data #analytics #datascience

As one of 5 companies selected as finalists for the Utah Tech Week Pitch Competition, we're excited to share our vision for and progress towards the future of Machine Learning with other innovators in tech. A huge thank you to the judges and the organizers of Utah Tech Week for their continual work in fostering and supporting innovation here in our home state of Utah. #data #technology #utah

ML researchers and practitioners have known for a while that deep learning often falls short with tabular data, despite recent advances. And recent research only confirms these limitations. But what do you do when XGBoost and other alternatives aren't working well either? This happens often when the patterns you're trying to learn are too complex, specifically highly non-linear patterns. The problem is that until now, there hasn't been a good ML solution for learning highly non-linear patterns on tabular data. This type of data occurs everywhere: in medicine, science, finance, industry and more. If you're interested in pushing beyond this limitation with a cutting-edge new approach to ML, reach out here or through https://lnkd.in/gBSPzFbb to learn more about joining the pilot program for AlephOne. #ml #machinelearning #data #ai #deeplearning

This was my prediction for the trajectory of ML & AI back in 2011: that advances made in hardware and compute power would dictate the direction of growth in the field until the branch that it had gone down reached its natural innovation asymptote. What dictates the nature of that asymptote? The structural limitations of the mathematical and conceptual foundations underpinning the technology. Every framework has them. And the way to overcome those limitations is to innovate at the level of the foundational mathematical framework itself. I think, and hope, that this leads to a renaissance of innovation and research at the theoretical core of ML and AI. As?Alan Mullenix, PhD?says, there's so much more meat on the bone. It's central to the technology we're building at Lucidity Sciences, and we're only getting started. #artificialintelligence #machinelearning #ai #ml #technology

Lucidity Sciences is accepting select organizations with machine learning problems in the scientific, industrial, and financial spaces for a pilot program for Aleph One. Aleph One is a foundationally new ML approach that exhibits significantly improved accuracy, efficiency, and performance on even the most complex problems on tabular data. If you're in one of these fields and have a problem that hasn't been satisfactorily solved by other modeling approaches, please contact us about exploring participation in our pilot program. Not sure if you qualify? We'll set up an exploratory call to help you identify and assess your ML use-cases and relevant data. #machinelearning #ML #analytics #data

Please join us on October 3rd at the Women Tech Awards in Salt Lake City!

If you want your data and machine learning initiatives to deliver meaningful impact, start with business strategy. Here's five strategic questions to help you identify opportunities for ML solutions that bring real business value. #machinelearning #data #datastrategy #strategy #ML #technology

All of those hundreds of parameters in your ML model might actually be sabotaging its performance: a closer look at overfitting and the meaning of simplicity.