Robotics Summer School 2022 - Interview with Simon Dufort-Labbé and Mathilde Hochedel
Vložit
- čas přidán 20. 08. 2024
- The 𝗥𝗼𝗯𝗼𝘁𝗶𝗰𝘀 𝗦𝘂𝗺𝗺𝗲𝗿 𝗦𝗰𝗵𝗼𝗼𝗹 2022 took place at Mila - Quebec AI Institute - from August 22 to 25.
The program consisted in lectures on robotics programming, simultaneous localization and mapping, or trajectory optimization, and hands-on practice on quadruped robots culminating into a challenge through an obstacle course.
The participating students united their brain power to compete in the final challenge, and had a lot of fun throughout the process.
Discover their interviews. They tell you about their discoveries, their difficulties and what they take away from this experience.
****************************************************
𝙄𝙣 𝙩𝙤𝙙𝙖𝙮'𝙨 𝙫𝙞𝙙𝙚𝙤
Simon Dufort-Labbé, PhD student at Mila working on optimal controls tools with the group of Pierre-Luc Bacon (mila.quebec/en....
Mathilde Hochedel, Field Application Engineer at Kinova (www.kinovarobo...)
Mila's Robotics Summer School 2022 participants.
****************************************************
𝗤𝘂𝗼𝘁𝗲𝘀
Simon: "𝘐 𝘭𝘦𝘢𝘳𝘯𝘦𝘥 𝘵𝘩𝘢𝘵 𝘳𝘰𝘣𝘰𝘵𝘪𝘤𝘴 𝘢𝘴 𝘢 𝘧𝘪𝘦𝘭𝘥 𝘰𝘷𝘦𝘳𝘢𝘭 𝘪𝘴 𝘭𝘦𝘷𝘦𝘳𝘢𝘨𝘪𝘯𝘨 𝘢 𝘭𝘰𝘵 𝘰𝘧 𝘵𝘰𝘰𝘭𝘴 𝘤𝘰𝘮𝘪𝘯𝘨 𝘧𝘳𝘰𝘮 𝘥𝘪𝘧𝘧𝘦𝘳𝘦𝘯𝘵 𝘱𝘭𝘢𝘤𝘦𝘴."
Mathilde: "𝘐𝘯 𝘢 𝘵𝘦𝘢𝘮, 𝘸𝘦 𝘭𝘦𝘢𝘳𝘯 𝘣𝘦𝘵𝘵𝘦𝘳, 𝘸𝘦 𝘢𝘳𝘦 𝘣𝘦𝘵𝘵𝘦𝘳, 𝘣𝘦𝘤𝘢𝘶𝘴𝘦 𝘸𝘦 𝘩𝘢𝘷𝘦 𝘥𝘪𝘧𝘧𝘦𝘳𝘦𝘯𝘵 𝘣𝘢𝘤𝘬𝘨𝘳𝘰𝘶𝘯𝘥𝘴, 𝘸𝘦 𝘩𝘢𝘷𝘦 𝘥𝘪𝘧𝘧𝘦𝘳𝘦𝘯𝘵 𝘴𝘬𝘪𝘭𝘭𝘴. 𝘞𝘦 𝘮𝘢𝘬𝘦 𝘢 𝘷𝘦𝘳𝘺 𝘨𝘳𝘦𝘢𝘵 𝘵𝘦𝘢𝘮. 𝘞𝘦 𝘭𝘦𝘢𝘳𝘯 𝘧𝘳𝘰𝘮 𝘦𝘢𝘤𝘩 𝘰𝘵𝘩𝘦𝘳, 𝘸𝘦 𝘭𝘪𝘴𝘵𝘦𝘯 𝘵𝘰 𝘦𝘢𝘤𝘩 𝘰𝘵𝘩𝘦𝘳, 𝘢𝘯𝘥 𝘵𝘩𝘢𝘵 𝘸𝘢𝘴 𝘵𝘩𝘦 𝘣𝘦𝘴𝘵 𝘱𝘢𝘳𝘵 𝘰𝘧 𝘵𝘩𝘦 𝘚𝘶𝘮𝘮𝘦𝘳 𝘚𝘤𝘩𝘰𝘰𝘭."
****************************************************
𝙉𝙤𝙩𝙚𝙨 𝙤𝙣 𝙩𝙝𝙚 𝙘𝙤𝙣𝙩𝙚𝙣𝙩
0'12": Pierre-Luc Bacon's research is at the intersection of reinforcement learning, optimal control and optimization. He's interested in developing novel algorithms and to challenge our theoretical understanding of reinforcement learning in the real world. His group works on end-to-end model-based reinforcement learning methods, meta-learning, and optimal control in continuous time.
0'14": 𝗥𝗟 = Reinforcement learning. It's an area of machine learning concerned with how 𝘪𝘯𝘵𝘦𝘭𝘭𝘪𝘨𝘦𝘯𝘵 𝘢𝘨𝘦𝘯𝘵𝘴* ought to take actions in an environment in order to maximize the notion of cumulative reward. Reinforcement learning is one of three basic machine learning paradigms, alongside supervised learning and unsupervised learning.
*In artificial intelligence, an 𝙞𝙣𝙩𝙚𝙡𝙡𝙞𝙜𝙚𝙣𝙩 𝙖𝙜𝙚𝙣𝙩 (IA) is anything which perceives its environment, takes actions autonomously in order to achieve goals, and may improve its performance with learning or may use knowledge.
0'20": 𝗢𝗽𝘁𝗶𝗺𝗮𝗹 𝗰𝗼𝗻𝘁𝗿𝗼𝗹 theory is a branch of mathematical optimization that deals with finding a control for a dynamical system over a period of time such that an objective function is optimized.
Like this video and subscribe for more great Robotics content!
#robotics #robots #quadrupedrobot #walkingrobot #machinelearning #deeplearning #reinforcementlearning #ai #artificialintelligence #education #research #universities #engineering #NVIDIAJetson #JetsonOrin #buildrobot #stellaquadruped #thinkahead