Intel AI

#intelAI worked with SueYeon Chung to apply her theory of manifold capacity to better understand speech recognition models such as DS2 [https://intel.ly/2xy7J2Z]. Authors also include Cory Stephens, Josh McDermott, Jenelle Feather, and Hanlin Tang.

#intelAI data scientists tackle the Voicebot Scaling Challenge with Kaldi ASR running on #XeonScalable processors to deliver incredible performance and maximize the value of existing large-scale production systems. https://intel.ly/2xSFpZb

#XeonScalable outperforms GPUs on NCF #inference. Neural Collaborative Filtering (NCF) leverages DL to improve the performance of recommender systems to predict a person's preferences based on other similar users. https://www.intel.ai/ncf-xeonscalable/ #intelAI

From medicine to manufacturing, #intelAI touches a multitude of industries. Intel IT’s Advanced Analytics team recently showcased some of its latest successes at AI Week hosted by Tel Aviv University: https://intel.ly/3a1r2jy

The DC Water bureau was able to reduce costs, optimize maintenance and improve customer service with #AI technologies powered by Wipro Limited and #intelAI. https://intel.ly/3cPl0nC

Intel, Dell, and the University of Florida researchers demonstrate how the large memory capacity of 2nd gen Intel Xeon Scalable CPUs allows researchers to build #AI models for brain tumor segmentation 3D imaging. https://intel.ly/36XJbw6 #intelAI

The Washington, D.C. water bureau was able to reduce costs, optimize maintenance and improve customer service with #AI technologies powered by Wipro and #intelAI. https://intel.ly/2RWZXHE

In this IEEE accepted paper, #intelAI researchers show how the resolution level of thermal imagery improves the accuracy of #DL for face recognition: https://intel.ly/2tu8jwW

#AI and HPC research project Etalumis proposes a new probabilistic inference approach to help run complex simulations in fields like high-energy physics: https://intel.ly/2ZiCjap #intelAI

How can we get autonomous systems to work together towards a common goal? A new framework from #intelAI researchers, MERL, demonstrates clear behaviors of team-formation and collaborative pursuit of an object. Learn more: https://intel.ly/35n1I58 #deep_RL

#intelAI researchers propose a novel 4-bit post-training quantization approach that may enable more capable, performant neural networks to fit into size-constrained inference hardware. Learn more: https://intel.ly/2PjlOWV

In this Venture Beat article, senior director of #intelAI lab Hanlin Tang explains how generative adversarial networks (GANs) are used to advance challenges like image upsampling and object transformation, by allowing model training with fewer data. https://intel.ly/2Qm8naz

Naveen Rao discusses the next wave of AI, increasing demand for AI compute, and how new #intelAI tech breaks barriers to meet the challenge: https://intel.ly/2PWYzDC