Privacy-Enhancing Technologies for Collaborative AI
Beschreibung
The student is expected to:
- Investigate new methods fore preserving privacy in distributed machine learning environments.
- Adapt libraries for secure multiparty computation to include novel primitives.
- Develop and test machine-learning use cases utilizing secure multiparty computation.
- Collaborate with multidisciplinary teams on exciting projects.
- Publish and present research findings at academic conferences.
- Contribute to intellectual property generation.
More information can be found here: Master Thesis Student in Privacy-Enhancing Technologies for Collaborative AI (m/f/d) - Huawei Research Center Germany & Austria
Voraussetzungen
- In-depth experience with machine learning and/or secure multiparty computation
- Proficiency in Python and C++
- Fluent written and spoken English language
Betreuer:
NN/LLM quantization optimization
Neural Networks (NNs), Large Language Models (LLMs), Quantization, Optimization
Beschreibung
With the rise of Large Language Models (LLMs) and Vision Transformers (ViTs), we investigate the return from expensive "Quantization-Aware Training" (QAT), towards smarter "Post-Training Quantization" (PTQ), where the choice of "what" to quantize and "how" is guided by (sometimes) sophisticated metrics. The student is expected to perform an evaluation of such metrics and their effect to the quantization results.
Voraussetzungen
Linear algebra (good to have)
Python coding skills (must)
Neural network basics (must)
Kontakt
If you are interested, please contact us by email and we can discuss more information and details. Also contact us if you have your own ideas that you would like to explore!