IEEE CIS Page IEEE SCV Page IEEE-Egrid IEEE-Egrid

On the Role of Structure in Learning for Robot Manipulation

Jeannette Bohg, Stanford University

Event Co-Sponsors:

Date & Time: Thursday, September 20, 2018, 6:00 PM – 8:00 PM PDT

Location: Intel SC12, 3600 Juliette Ln, Santa Clara, CA 95054

Directions: Intel-SC12-Auditorium.pdf

Registration Link: (Mandatory) : EventBrite Link

Registration Fee: IEEE CIS/RAS/VTS members: free
         Students - free
         IEEE members - free
         Non-members - $10 (Register at Door $15)

Share Event with your Colleagues:

LinkedIn Email Facebook Twitter
prof_JeannetteBohg

Abstract: Recent approaches in robotics follow the insight that perception is facilitated by interaction with the environment. First, this creates a rich sensory signal that would otherwise not be present. Second, knowledge of the sensory dynamics upon interaction allows prediction and decision-making over a longer time horizon. To exploit these benefits of Interactive Perception for capable robotic manipulation, a robot requires both: methods for processing rich, sensory feedback and feedforward predictors of the effect of physical interaction. In the first part of this talk, I will present a method for motion-based segmentation of an unknown number of simultaneously moving objects. The underlying model estimates dense, per-pixel scene flow that is then followed by clustering in motion trajectory space. We show how this outperforms state-of-the-art in scene flow estimation and multi-object segmentation. In the second part, I will present a method for predicting the effect of physical interaction with objects in the environment. The underlying model combines an analytical physics model and a learned perception part. In extensive experiments, we show how this hybrid model outperforms purely learned models in terms of generalisation. In both projects, we found that introducing structure greatly reduces training data, eases learning and provides extrapolation. Based on these findings, I will discuss the role of structure in learning for robot manipulation.

Biography: Jeannette Bohg is an Assistant Professor of Computer Science at Stanford University. She was a group leader at MPI until September 2017 and remains affiliated as a guest researcher. Her research focuses on perception for autonomous robotic manipulation and grasping. She is specifically interesting in developing methods that are goal-directed, real-time and multi-modal such that they can provide meaningful feedback for execution and learning. Before joining the Autonomous Motion lab in January 2012, Jeannette Bohg was a PhD student at the Computer Vision and Active Perception lab (CVAP) at KTH in Stockholm. Her thesis on Multi-modal scene understanding for Robotic Grasping was performed under the supervision of Prof. Danica Kragic. She studied at Chalmers in Gothenburg and at the Technical University in Dresden where she received her Master in Art and Technology and her Diploma in Computer Science, respectively.

IEEE CASS-SV Artificial Intelligence For Industry Forum

Event Co-Sponsors:

  • IEEE Circuits and Systems Society Santa Clara Valley Chapter
  • IEEE Communications Society Santa Clara Valley Chapter
  • IEEE Computational Intelligence Society Santa Clara Valley Chapter
  • IEEE Computer Society Technical Committee on Multimedia Computing
  • IEEE Signal Processing Society Santa Clara Valley Chapter
  • Tau Beta Pi San Francisco Bay Area Alumni Chapter
  • IEEE Computer Society Technical Committee on Semantic Computing

Date & Time: Fri, September 21, 2018 1:00 PM – 5:00 PM PDT

Location: Intel Santa Clara SC12 auditorium, 3600 Juliette Ln, Santa Clara, CA 95054

Directions: Intel-SC12-Auditorium.pdf

Registration Link: (Mandatory) : EventBrite Link

Registration Fee: Free

Share Event with your Colleagues:

LinkedIn Email Facebook Twitter

List of Topics: Algorithm-architecture co-design for energy-efficient deep learning, including algorithm optimization (e.g., novel numerical representation, network pruning/compression) and accelerator architectures (e.g., programmable SoC).

List of Speakers:

  • Dr. Debbie Marr (Intel)
  • Prof. Vivienne Sze (MIT)
  • Dr. Mark Sandler (Google)
  • Dr. Jongsoo Park (Facebook)

IEEE Workshops on Machine Learning, Convolutional Neural Networks, and Tensorflow

Dr. Kiran Gunnam, IEEE Distinguished Speaker and Distinguished Engineer - Machine Learning & Computer Vision at Western Digital

Event Co-Sponsors:

  • IEEE Silicon Valley Chapters of ComSoC, ITSoC, CIS
  • Apollo AI [Disrupting the automotive industry with groundbreaking solutions. Building low-cost super-safe self-driving system and HD maps]

Date & Time: Monday September 24 & Tuesday september 25, 4:00 PM – 9:00 PM PDT

Location: Texas Instruments Building E Conference Center, 2900 Semiconductor Drive, Santa Clara, CA 95051

Directions: Texas_Instruments_BuildingE_ConferenceCenter.pdf

Registration Link: (Mandatory) : EventBrite Link

Registration Fee: IEEE CIS/RAS/VTS members: free
       TI Employee: Workshop I and II: $45; Workshop I Only: $27; Workshop II (Only): $27
       IEEE members: Workshop I and II: $270; Workshop I Only: $180; Workshop II (Only): $180
       IEEE ComsSoC,ITSoC,CIS members: Workshop I and II: $225; Workshop I(Only):$157.50;Workshop II(Only):$157.50
       Non-members: Workshop I and II: $315; Workshop I Only: $202.50; Workshop II (Only): $202.50

Share Event with your Colleagues:

LinkedIn Email Facebook Twitter

List of Topics: For the detailed list of topics covered, please see the link. Course slides in PDF and other workshop materials will be shared with registered attendees 5-days before the course. In addition, workshop materials with Tensorflow installation are provided also as docker image to have a worry free setup. Attendees for workshop 2 should bring their own laptop prepared with provided docker image or Tensorflow+provided examples.

Target Audience: Engineers, researchers, practitioners and students who are interested in machine learning, convolutional neural networks, recurrent neural networks, reinforcement learning and their implementations on GPUs and FPGAs. This workshop series will particularly benefit people who intend to develop machine learning techniques and applications that can keep improving themselves after seeing more and diverse data to achieve intelligence. Key words: Machine Learning, Deep Learning, Supervised Learning, Unsupervised learning, CNN, RNN, GPU, FPGA Prerequisite Knowledge: Basic knowledge of matrices, vectors, derivatives, probability. You may also want to review these useful guides.

Biography: Dr. Kiran Gunnam is an innovative technology leader with vision and passion who effectively connects with individuals and groups. Dr. Gunnam's breakthrough contributions are in the areas of advanced error correction systems, storage class memory systems and computer vision based localization & navigation systems. He has helped drive organizations to become industry leaders through ground-breaking technologies. Dr. Gunnam has 75 issued patents and 100+ patent applications/invention disclosures on algorithms, architectures and real-time low-cost implementations for computing, storage and computer vision systems. He is the lead inventor/sole inventor for 90% of them. Dr. Gunnam’s patented work has been already incorporated in more than 2 billion data storage and WiFi chips and is set to continue to be incorporated in more than 500 million chips per year. Dr. Gunnam is also a key contributor to the precise localization and navigation technology commercialized for autonomous aerial refueling and space docking applications. His recent patent pending inventions on low-complexity simultaneous localization and mapping (SLAM) and 3D convolutional neural network (CNN) for object detection, tracking and classification are being commercialized for LiDAR+camera based perception for autonomous driving and robotic systems. Dr. Gunnam received his MSEE and PhD in Computer Engineering from Texas A&M University, College Station. He is world-renowned for balance between strong analytical ability and pragmatic insight into implementation of advanced technology. He served as IEEE Distinguished Speaker and Plenary Speaker for 25+ events and international conferences and more than 3000 attendees in USA, Canada and Asia benefited from his lecture talks. He also teaches graduate level course focused on machine learning systems at Santa Clara University.