/Probabilistic bit enabled Machine Intelligence for edge devices

Probabilistic bit enabled Machine Intelligence for edge devices

PhD - Leuven | More than two weeks ago

Algorithm and system co-optimization for efficient probabilistic computing on edge devices.

Machine learning recently enables to tackle more and more problems such as cognitive recognition, robotics, discovery, and many others. The machine learning algorithms, however, operate in a deterministic way, unable to handle probabilistic inputs or quantify uncertainty associated with their prediction, resulting in false positive responses. One way to increase confidence about the results, is the use of giant data sets and very large-scale models, both unachievable in the context of edge devices. 


The alternative of probabilistic machine learning models, such as Bayesian Neural Networks, or Probabilistic Graphical Models with Gibbs sampling, allow to incorporate uncertainty and train with a limited data set. This offers an attractive model for inference and training in resource constrained devices, such as smart phones, glasses or other embedded devices. Yet, probabilistic network learning and intelligence depend crucially on the careful probabilistic representation of uncertainty. This requires the device to be capable of generating tuneable stochastic samples, at a very high throughput and with low energy and area footprint. Current digital CMOS implementations, however, fail to offer such solutions.  


On-chip probabilistic samples can also be implemented by exploiting emerging silicon devices, such as magnetic nanodevices (called “p-bits”). These p-bits promise to be more energy efficient and compact and allow to conceptualize a system exploiting the combination of many magnetic nanodevices, each capable of generating tuneable random bits at GHz speeds. The outputs of these devices can be smartly combined in an analog or digital fashion, to generate a wide variety of tuneable probabilistic distributions. For foresee such building blocks to become a key block in the probabilistic machine learning processors of the future. 


In this PhD, the student will be responsible to explore the optimal design of p-bit-based random number generators, as well as their integration into machine learning processor chips for probabilistic algorithms.  


The thesis work will involve the following task: 

  1. Understanding of probabilistic algorithms, e.g., quantum Montecarlo for edge computing and data fusion applications. 

  2. Understanding of mechanism of probabilistic elements. 

  3. Designing of probabilistic bit-based random number generators and benchmarking their performance with different state of art digital designs. 

  4. Integrate the random number generators into a machine learning processor. 

  5. Efficiency benchmarking of the resulting processor for giving probabilistic algorithms. 

  6. Experimental verification of the magnetic random access memory enabled processor hardware. 


This thesis work allows the student to learn the probabilistic algorithm, probabilistic bit-based machine learning, co-optimization of probabilistic bit-based ML performance, and testing of the hardware of the building blocks

Required background: MTech. Electronics/Electrical Engineering, MTech. AI and ML.

Type of work: 90% modeling, design, testing, and 10% literature.

Supervisor: Marian Verhelst

Daily advisor: Ankit Kumar, Marian Verhelst

The reference code for this position is 2024-035. Mention this reference code on your application form.

Who we are
Accept marketing-cookies to view this content.
Cookie settings
imec's cleanroom
Accept marketing-cookies to view this content.
Cookie settings

Related jobs

Staff and Operations Manager

Within Specialty Components and Platforms Development (SCPD) unit, we are looking for a motivated and enthusiastic colleague to lead staff and operational activities. Currently this position is mapped at manager level with the outlook of growing further in the organization.

Senior R&D Engineer: Electronic System Architecture

We are looking for an enthusiastic person eager to take ownership during the whole electronic system design and development cycle.

Layout and tape-out engineer DST/SCPD

As tape-out and lay-out engineer, you will play a key role to support IC manufacturing in advanced technologies for logic scaling, Si Photonics, Imagers, Life Science, GaN Power and other specialties.

Real-time monitoring of the remaining useful life of an electronic system with a data-based approach

The aim of this PhD track is to develop an AI methodology, using machine learning (ML) approaches, to predict the remaining useful lifetime of electronic systems by fusing data obtained from both sensors and simulations.

Senior Analog and Mixed-Signal Design Engineer for Sense & Actuate SoCs

Design high-performance analog and mixed signal circuits for sensor and actuator interfaces. Develop complex Systems on Chip from concept to low volume production. Join us and find the perfect balance between research and industry.

Program Director/Manager - Product research and development for Specialty Integrated Circuits - PDT

Strengthen imec’s offering for niche Product development and production and to contribute to the strategy definition of the group.
Job opportunities

Send this job to your email