Connectionists: Exciting job opportunities in Spiking Neural Networks for Machine Learning

[Arindam Basu]

Dear all

Several exciting positions available at School of EEE, NTU in industry sponsored projects for developing low-power, neuromorphic machine learning chips, embedded platforms and supporting algorithms.

Established in 1981, the SCHOOL OF ELECTRICAL AND ELECTRONIC ENGINEERING (EEE) is one of the founding Schools of the Nanyang Technological University. Built on a culture of excellence, the School is renowned for its high academic standards and strong tradition in research. To support teaching and cutting-edge research, EEE is host to 11 research centres and more than 50 laboratories, which are well-equipped with modern facilities and state-of-the-art equipment. With about 200 faculty members and an enrolment of more than 4,000, of which about 1,300 are graduate students, it is one of the largest EEE schools in the world.

The positions are in the BRAIN lab (http://www3.ntu.edu.sg/home/arindam.basu/ ) under the supervision of Dr. Arindam Basu. Selected candidates will get to work with cutting edge neuromorphic image sensors and some of the lowest power machine learning chips in the world such as the IBM True North and the ELM IC developed in BRAIN lab. It has been widely covered in media for biomedical applications (http://www.channelnewsasia.com/news/singapore/ntu-scientists-developing/2506294.html ) and the corresponding publication is one of the most popular article in IEEE TBCAS since publication (http://ieeexplore.ieee.org/xpl/topAccessedArticles.jsp?punumber=4156126 ).

The candidate will get the chance to use this ELM IC for a new application (object detection in spike based cameras for unattended ground sensors) and develop new generations of this design with ample opportunity for basic research as well as commercialization through startup. The candidates also have a chance to work with IBM’s True North technology of ultra low-power spiking neural networks. Details of the positions are as follows:

Project Name: Unattended Ground Sensors
All jobs in this project will typically have a 1 year initial contract followed by extension up to 4 years based on satisfactory performance. Remunerations will be competitive to world standards and be negotiable depending on candidate’s relevant experience.

1. Job Title: Research Fellow/Research Associate
Job Field: Machine learning, Neural Networks

Educational Requirements: PhD (for Postdoctoral Research Fellow) or M.S + >2 yrs of experience (for Research Assistant) in related topic

Job description: Develop algorithms for object detection and classification from asynchronous event based camera outputs. The candidate has the chance to work with cutting-edge low-power machine learning  hardware  such as the NTU ELM IC and IBM True North. Selected candidate has to:

* Train deep networks for IBM TrueNorth based deployment
* Develop randomized neural network algorithms (such as ELM, RVFL) to classify objects
* Optimize neural networks for given hardware constraints
* Develop strategies for fusion of acoustic and camera inputs to reduce false positives
* Develop strategies to partition tasks between several neuromorphic processors and general purpose processors like ARM/MCU or FPGA.

Job requirement:
Must have:
* Research work related to field of neural network, machine learning
* Knowledge of MATLAB coding including usage of and toolbox for fixed point arithmetic and neural network
* Knowledge of Spiking neural networks

Good to have:
* Experience in MCU coding
* Experience in working with event based sensors


2. Job Title: Research Assistant/ Project Officer
Job Field: Embedded hardware

Educational Requirements: M.S (for Research Assistant) or B.S + >2 yrs of experience (for Project Officer) in ECE/ EE/CSE

Job description: Selected candidate has to help in system integration of custom machine learning ELM IC with neuromorphic camera and FPGA platform.

* Main job is to write Verilog/VHDL codes for interfacing FPGA with custom ELM IC developed by team and neuromorphic image sensor and IBM TrueNorth
* Develop library of Verilog/VHDL codes for basic image processing blocks
* Need to co-work with algorithm person to optimize the neural network for hardware friendly implementation

Job requirement:
Must have:
* Experience in HDL coding, using FPGA kits and interfacing FPGA with computer and other components
* Knowledge of debugging using logic analyser
* Knowledge of at least one of MATLAB/OPENCV/Python programming language for interfacing with FPGA kit

Good to have:
Knowledge of Neural Networks and Image processing algorithms



Regards,
Arindam
________________________________
CONFIDENTIALITY: This email is intended solely for the person(s) named and may be confidential and/or privileged. If you are not the intended recipient, please delete it, notify us and do not copy, use, or disclose its contents.
Towards a sustainable earth: Print only when necessary. Thank you.