Jensen Huang (Nvidia Co-founder) – NVIDIA CEO Press Meeting Tel Aviv 2017 (2017)


Chapters

00:00:00 Machine Learning Revolution: Deep Learning's Transformative Impact
00:02:39 The Emergence of GPU-Accelerated Computing and Its Impact on Unsolvable Problems
00:10:18 AI and GPU Innovations
00:16:48 The Rise of Autonomous Vehicles: NVIDIA's Journey and Industry Challenges
00:27:37 Navigating the Technological Frontier: Challenges and Opportunities in AI and Computing
00:37:21 Artificial Intelligence and the Future of Transportation
00:41:14 NVIDIA's Strategies for Future Investment and Collaboration
00:49:24 AI and Autonomous Cars: The Future of Education and Transportation

Abstract

Deep Learning Revolution and NVIDIA’s Role in Shaping the Future

The Pioneering Leap of Deep Learning in AI and Computing

Deep learning, a breakthrough in machine learning, involves self-writing software and self-learning computers. It can find important features from a large number of examples without explicit instructions, which has made it almost magical. Deep learning utilizes convolutional neural networks to detect features, starting from low-level features and gradually building higher-level abstract features. This allows the network to identify intricate details and even differentiate between species of cats. Deep learning models are robust to changes and can generalize learned knowledge. Even if the color of an object is altered, the model can still accurately identify it, demonstrating its ability to recognize subtle details while being difficult to deceive.

Deep learning is not limited to computer vision; it has been applied to various other domains. It is particularly effective in analyzing messy data when traditional software development approaches may be impractical. By training on numerous examples, deep learning can automatically determine features and make predictions, classifications, and recommendations based on the data.

GPU-Accelerated Computing: Catalyzing the Deep Learning Epoch

Jensen Huang, CEO of NVIDIA, has emphasized the critical role of GPU-accelerated computing in meeting the intensive computational demands of deep learning. GPUs have surpassed traditional CPU performance, primarily due to their ability to accelerate specific program kernels, achieving speed enhancements of up to 50 times. This leap in computational capability has been instrumental in advancing AI and deep learning, enabling the exploration of intelligence in ways previously deemed impossible.

Jensen Huang highlights the growing exploration of artificial intelligence capabilities using deep learning. The stagnant performance of CPUs due to limitations in transistor performance and architectural advancements has created a demand for more computational capabilities. NVIDIA pioneered GPU-accelerated computing to intercept and offload key kernels of programs, enabling significant acceleration. The speedups achieved by GPUs are comparable to decades of progress under Moore’s Law. This acceleration has made it possible to solve problems that would have otherwise taken much longer, effectively acting as a time machine. The tipping point in GPU adoption has led to widespread application in various fields. Previously unsolvable problems in sectors such as transportation, healthcare, finance, enterprise, farming, retail, and consumer applications are now being addressed through GPU acceleration.

NVIDIA’s Strategic Vision in Automotive and AI

NVIDIA’s journey in the automotive industry, initially focused on infotainment systems, has evolved towards a concentration on AI and autonomous driving. Huang envisions Level 5 autonomous taxis by 2020, a prediction underlining NVIDIA’s significant anticipated market share in the autonomous vehicle hardware segment. Furthermore, the company’s ambition extends to reshaping traditional infotainment systems through intelligent agents, a move that mirrors its commitment to integrating cutting-edge technology with conventional automotive engineering.

NVIDIA aims to reshape infotainment by transforming it into an AI-powered agent to enhance the user experience. The automotive industry’s cautious approach to innovation is due to the significant impact of new technologies on society and the need for meticulous engineering. NVIDIA emphasizes the importance of balancing traditional engineering practices with the integration of groundbreaking AI technology to revolutionize vehicles. Autonomous vehicles are expected to hit the roads sooner than initially anticipated, with level 5 taxis projected to emerge in 2020 and level 4 autonomous vehicles by 2021-2022. NVIDIA anticipates a substantial market share in the autonomous vehicle industry due to its focus on level 4 and level 5 cars and its dedicated efforts in this field for almost a decade. The company believes that the market is vast, encompassing various types of vehicles and applications, and it’s too early to focus on specific market share percentages. NVIDIA prioritizes developing and refining the technology, ensuring performance, safety, and assisting partners in bringing autonomous vehicles to the market. NVIDIA acknowledges the success of Mobileye and Intel in the autonomous vehicle space. It recognizes Mobileye’s expertise in ADAS technology and anticipates strong competition from the company in the autonomous vehicle domain. Intel’s capabilities and determination in pursuing new ventures are acknowledged, and NVIDIA views this as a competitive challenge. NVIDIA emphasizes its unique strengths in creating an open platform with a fully driveable stack, positioning it well in the autonomous vehicle landscape.

Robot Taxis and Autonomous Cars:

NVIDIA expects robot taxis to operate relatively soon in geo-fenced areas with complete confidence in their maps and surroundings. These areas could be large cities like San Francisco, New York City, or Phoenix, which have been extensively mapped digitally. Robot taxis are equipped with numerous sensors that record everything, providing valuable data for accident analysis and fault determination. Determining the regions where these technologies will emerge first is challenging, but the benefits of robot taxis are most significant in urban environments with a high demand for taxi services.

NVIDIA’s Ethical Stance and Open Computing Philosophy

In a world increasingly reliant on AI, NVIDIA acknowledges the dual nature of this powerful technology. Huang stresses the importance of maintaining an open dialogue and awareness to mitigate potential harms. By keeping its computing platform open and supporting diverse frameworks and cloud providers, NVIDIA aligns itself with a philosophy that champions the inherent goodness of humanity and the democratization of powerful technologies.

Transforming Transportation: The Vision of Autonomous and Electric Vehicles

NVIDIA’s vision extends to transforming the automotive industry with autonomous and electric vehicles (EVs). The immediate goal is to enhance the driving experience and accelerate the transition to EVs and autonomous vehicles (AVs), while the long-term vision focuses on reducing the number of cars on the road and promoting shared transportation models for societal betterment.

NVIDIA has been involved in the automotive industry for 10 years, initially focusing on infotainment and now shifting towards artificial intelligence and autonomous driving. NVIDIA emphasizes the importance of balancing traditional engineering practices with the integration of groundbreaking AI technology to revolutionize vehicles.

NVIDIA’s Investment Strategy and Collaboration with Mellanox

NVIDIA’s shift from a passive to an active investment approach highlights its intention to accelerate promising startups using its technology. The collaboration with Mellanox in developing GPU direct technology for efficient data transfer exemplifies NVIDIA’s strategic partnerships aimed at optimizing performance and advancing its technological capabilities.

Collaboration with Startups:

NVIDIA actively invests in promising startups that align with its mission of creating groundbreaking ideas. The company’s goal is to accelerate the growth of these startups and bring their products to market faster. NVIDIA offers unique assistance through its technology platform, technical capabilities, and marketing exposure.

Mellanox Collaboration:

NVIDIA has partnered with Mellanox to address the challenges of data transfer for high-speed computation. The collaboration resulted in the development of GPU direct, a software and architecture solution that allows data from Mellanox NICs to directly enter NVIDIA GPUs. GPU direct optimizes data transfer and enhances the performance of NVIDIA’s GPU computing platform.

Conclusion

In summary, NVIDIA, under Jensen Huang’s leadership, is navigating a transformative era in computing and AI. From pioneering developments in deep learning to strategic expansions in the automotive industry and global collaborations, NVIDIA is at the forefront of shaping a future where AI and GPU-accelerated computing play central roles. The company’s commitment to ethical considerations, open platforms, and strategic investments further cements its position as a key player in the evolving landscape of technology and innovation.


Notes by: Rogue_Atom