50 Most Innovative Companies to Watch 2021
Today’s artificial intelligence solutions tend to be specialized in specific applications and functions, and rely on maintenance by a select few developers. They are talent-intensive, expensive, difficult to reproduce, and overall unsustainable. As a result, AI currently remains inaccessible to most organizations. Companies are on a constant quest to improve yield, improve quality and lower costs while meeting sustainability goals. With increasing digitization, more and more data is now available across all industrial sectors, but the sheer volume and variety of sensor data make it hard to harness practically. Taking advantage of the linear, non-linear, long range, temporal or difficult-to-predict patterns in your data is challenging without an array of AI techniques at your disposal.
Petuum is one such company that provides an AI software platform that allows enterprises to design, build, experiment, customize, operate and own vertical AI solutions in a wide range of industries and areas, such as healthcare, industrial manufacturing and utilities, financial services, telecommunications, autonomous driving, and beyond. Petuum lets enterprises easily understand and leverage AI to gain deep insight for better decision-making and improved productivity and efficiency. Petuum Industrial AI delivers sophisticated data processing capabilities, robust deep learning and machine learning techniques, and standardized integrations with operational control systems for solutions that learn and adapt continuously and offer a multi-layered, multi-faceted understanding of complex industrial systems. The industrial AI integrates with existing plant and corporate data infrastructures such as OSIsoft PI systems and control systems to directly deliver timely parameter changes, so that completely automated, yet supervised operations can become a reality.
The Petuum Advantage
Achieve the Impossible: The company’s award-winning scientists help you unlock the transformative benefits of artificial intelligence. Its expertise in tackling the most complex problems helps you find answers to the toughest challenges in the shortest period of time.
Get AI Faster: Harness any data- tabular, unstructured, semi-structured, images, video, audio, streaming to implement practical AI solutions. Recraft, recombine, and reuse datasets and models on the fly. Drop in, join and scale any model with support for commonly used frameworks.
Reimagine Your Business: Leapfrog your competition with advanced AI capabilities that drive growth, create new opportunities and position your business for the future.
Revolutionary AI-Based Products Offered
BlinkLabel: Petuum’s advanced AI-enabled data labeling and annotation software offers reliable, repeatable, blazing fast data labeling with a high degree of accuracy that continuously improves over time. Reduce your annotation time significantly with its interactive and adaptive software. This product detects and segments objects on an image based on a predefined set of classes and assigns a label instantaneously. No clicks are required, although users are still able to refine if necessary. Applied ML/DL models that learn based on real data, avoiding assumptions of linearity or correlation, and sensitive to linear, non-linear, temporal and long-range patterns in the data
CASL: It provides a unified toolkit for composable, automatic, and scalable machine learning systems, including distributed training, resource-adaptive scheduling, hyperparameter tuning, and compositional model construction. Gain increased throughput, improved yield, and greater productivity while reducing emissions and enhancing energy efficiency with artificial intelligence. It provides sophisticated techniques for data cleaning and normalizing so the AI is robust to noisiness, unlike traditional solutions. Repeatable machine learning pipelines and standardized interfaces with control systems, for scalable, rapid deployment and time to value
The Leader Upfront
Aurick Qiao is the Chief Executive Officer of Petuum. He is also the Heard of Engineering at Petuum. He completed his Ph.D. work at the Computer Science Department at Carnegie Mellon University, where he researched resource adaptability and elasticity for distributed machine learning systems. Before Petuum and CMU, Aurick studied computer science and combinatorics and optimization at the University of Waterloo, where he also researched multi-pivot quicksort. He has also been successful in competitive programming.