Jarad Forristal

Title: The Innovations of Jarad Forristal

Jarad Forristal is an accomplished inventor based in Marfa, TX (US). He has made significant contributions to the field of deep learning and artificial intelligence. His innovative approach to training neural networks has garnered attention in the tech community.

Latest Patents

Jarad Forristal holds a patent for a "Deep learning model training system." This invention involves a computer that trains a neural network. The process includes executing a neural network with a weight vector to compute a gradient vector using a batch of observation vectors. Eigenvalues are computed from a Hessian approximation matrix, and a regularization parameter value is determined using the gradient vector, eigenvalues, and a step-size value. The invention also computes a search direction vector, reduction ratio value, and updated weight vector, along with an updated Hessian approximation matrix. The step-size value is then updated using the search direction vector.

Career Highlights

Jarad Forristal is currently employed at SAS Institute Inc., where he continues to develop innovative solutions in the field of artificial intelligence. His work focuses on enhancing the efficiency and effectiveness of deep learning models.

Collaborations

Some of Jarad's notable coworkers include Joshua David Griffin and Seyedalireza Yektamaram. Their collaborative efforts contribute to the advancement of technology in their respective fields.

Conclusion

Jarad Forristal's contributions to deep learning and neural network training exemplify the spirit of innovation. His patent and ongoing work at SAS Institute Inc. highlight his commitment to advancing technology.