Master thesis proposals - external

SkyCraft AB

Master projects in CV and ML on powerline detection

Description

Viscando AB

CV and ML diploma projects at Viscando

Description

Arkus AI

Apply Machine Learning and Computer Vision in Genetic Diagnostics

Arkus AI is a fast-growing AI (Artificial Intelligence) startup created in January 2020, as a result of an industry project co-founded by a Vinnova innovation grant for SME. Arkus AI is created with an ambition to innovate AI technologies in the field of genetic diagnostics and healthcare. Its vision is to deliver AI worldwide in the spirit of continuous innovation and earnest compassion. We are on track to deliver the first AI-powered automation tool in clinical genetics and we are delighted to offer the following two Master’s thesis projects.

Project I: Research and develop an innovative solution that combines both deep learning and computer vision technics to understand medical images from clinical genetics. For this project, the student must have an outstanding competency in computer vision, in addition to deep learning skills.

Project II: Innovate in deep learning technics in the search of a high performing classifier of processed medical images from clinical genetics. For this project, the student must be fluent with existing deep learning methods in the field.

Required competencies: Passionate for innovation, self-motivated, and can work independently; Competent in machine learning in particular deep learning; Fluent in Python programming language; Some industrial experience in software programming; Excellent in oral and written English.

Students will be mentored and challenged by Arkus’ exceptionally talented and compassionate engineers. Upon completion of the thesis, highly qualified students will be offered positions to walk the startup journey with us.

Location: Working remotely until Covid situation clears. Regular meet-ups at AI Sweden’s Stockholm office or THINGS at KTH. No requirement to be physically in Stockholm

Contact: join@arkus.ai or ying@arkus.ai, Ying Cheng/Co-Founder, M: 0760901386

Babyshop Group

Leveraging transfer learning for multi-label attribute prediction in fashion images, using deep learning

At Babyshop Group, there exist tens of thousands of images of children’s apparel. Currently, new items are labeled by hand, where each item has multiple attributes, eg. color, category, pattern, neckline, and style, etc. Since the workload of this manual task is immense, a need for a support system is high. In literature, a multitude of previous work has been done in regards to multi-label attribute prediction on fashion clothing, most notably DeepFashion and iMaterialist Fashion. Here, deep learning models have been trained on hundreds of thousands of labeled images to predict corresponding attributes, which has proven highly successful. However, almost all of these images are of adult fashion, with very little emphasis on childrens’ fashion. The question then arises, is it possible to combine the pre-trained models from previously mentioned multi-label attribution prediction models that are trained on almost exclusively adults’ fashion, together with the learnings from transfer-learning frameworks - in order to successfully perform multi-label attribution prediction on childrens’ fashion?

Required qualifications: MSc studies in Computer Science, Machine Learning, Computer Engineering, Mathematics, Physics, or related field; Good understanding of machine learning frameworks such as Keras, TensorFlow, PyTorch, Scikit-Learn, and/or Spark; Proficiency in Python and Git; Knowledge about data wrangling and data munging, using SQL, Pandas, and Numpy.

Web version of the proposal with instructions on how to apply.

Time frame: from early January to mid-June.

Contact: Marcus Svensson (marcus.svensson@babyshop.se), Data Scientist at Babyshop Group

Scania Autonomous Transport Solutions

A thesis project at Scania is an excellent way of making contacts for your future working life. Many of our current employees started their career with a thesis project. 
The concept of autonomous driving has rapidly transitioned from being a futuristic vision of robotics and artificial intelligence to being a present-day reality. Scania is among the companies aiming to sell fully autonomous vehicles a few years from now, but many research challenges still needs to be solved.

Within the topics of robotics, perception and learning we have several interesting projects:

• Risk-aware Decision Making with POMDPs
• Online Estimation for of Safety Margins for Risk Assessments 
• Behavior Prediction of Surrounding Cyclists for Autonomous Driving
• Collision checking for articulated buses
• Steering Methods for Nonholonomic Motion Planning

More projects may also be published at Scania Career.

Contact:
See link for each project, or contact Truls Nyberg, Industrial PhD Student Scania/RPL
trulsny@kth.se

Ericsson

In an automated warehouse where autonomous robots load trucks with products while sharing the same environment with humans, a proper safety analysis is performed to avoid the hazardous situations without compromising the productivity. We have a basic safety analysis mechanism using image processing for object identification and risk assessment and mitigation (using fuzzy logic, RL approaches) to provide safety. The tasks of this thesis will be to extend the safety model: (1) adding safety from communication perspective i.e. among robots and between the robot and the edge / cloud; (2) performing highly-computational processing in edge / cloud; (3) looking into the Explainability of the used AI (XAI) techniques for Trustworthiness perspective; (4) model compression to run complex deep learning architectures in devices with limited hardware.
The high-level task is to perform safety analysis and image processing in the cloud and locally at the attached hardware and comparing both with respect to safety, performance, energy usage in human-robot collaborative use case.

Required qualifications:  MSc studies in Computer Science, Electrical and Computer Engineering or similar area; Excellent programming skills in C/C++, or Python or Java or Matlab; Good knowledge of concepts in machine learning (e.g. deep learning), robotics, ROS, etc.; Experiences with machine learning libraries Tensor flow, Keras, PyTorch, sci-kit learn etc.; Knowledge of safety analysis techniques, XAI, communication protocols and technologies (edge/cloud, D2D) is a bonus; Like to build end to end prototypes and concepts.

Contact: Rafia Inam (rafia.inam@ericsson.com), Alberto Hata (alberto.hata@ericsson.com )

Full project description here 

H&M

These projects will be performed in collaboration with H&M and aim to provide efficient solutions for some of the problems that the company is facing. The successful applicant will have an opportunity to closely interact with experts from H&M, have access to the real-world data owned by the company, and receive guidance from KTH researchers.

Project 1: Quality prediction

Every day, H&M orders, transports, and sells thousands of items. Quality control is an important part of this process. When boxes with goods are delivered to one of the warehouses, random quality checks are performed. There are multiple factors that can potentially affect the quality of the products in a specific box: they come different suppliers overseas, are transported in different ways (by ships, trains, or planes), and indirect factors such as weather also may play a certain role.

In this project, we aim to develop a probabilistic model that will predict the quality of the products based on the information about the supplier, the means of transportation, etc. To train the model, H&M will provide the relevant data. If implemented successfully, the system could later be used to identify the boxes that are more likely to require quality control.

Required qualifications: proficiency in Machine Leaning and Data Science. Applicants are expected to have passed KTH courses such as Advanced Machine Learning, Project Course in Data Science, Probabilistic Graphical Models, or equivalent.

Contact person: Anastasiia Varava (KTH) varava@kth.se

Project 2: Carton fillrate optimization

To transport the products, H&M needs to make decisions on how to package them. Depending on the characteristics of the products being packed (type of product, size, material, fragility, etc), different types of cartons can be used, and different number of items are packed into each of them.

Previously, this decisions were made by humans on a case-to-case basis.

In this project, we aim to automate this process. Based on the data collected earlier at H&M (item descriptions and packaging specifications), we aim to develop a model that will find the best way to package novel products. The model will need to capture specific features of the products that can potentially affect these decisions, and thus can be trained on existing data that will be provided by the company.

Required qualifications: proficiency in Machine Leaning and Data Science. Applicants are expected to have passed KTH courses such as Advanced Machine Learning, Project Course in Data Science, Probabilistic Graphical Models, or equivalent.

Contact person: Anastasiia Varava (KTH) varava@kth.se

Project 3: Inbound leadtime prediction

H&M orders large amounts of goods from multiple different suppliers in various countries to markets all over the world. Each product is expected to be delivered by a specific date. However, delays in production and transportation are inevitable, especially when the products are transported by ship due to multiple factors such as weather conditions, scheduling imprecision, multiple connections between different routes, etc. The goal of this project is to develop a probabilistic model that will be trained on the historical data provided by the company to predict the expected time needed for the goods to be manufactured and delivered to markets.

Required qualifications: proficiency in Machine Leaning and Data Science. Applicants are expected to have passed KTH courses such as Advanced Machine Learning, Project Course in Data Science, Probabilistic Graphical Models, or equivalent.

Contact person: Anastasiia Varava (KTH) varava@kth.se

Reinforcement Learning at Flexible Alternating Current Transmission Systems

You will be part of the Business Unit Grid Integration, located in Västerås. FACTS (Flexible Alternating Current Transmission Systems) technologies provide more power and control in existing AC as well as green-field networks and have minimal environmental impact. With a complete portfolio and in-house manufacturing of key components, ABB is a reliable partner in shaping the grid of the future. Please find out more about our world leading technology at www.abb.com/facts.

Problem description

At FACTS you will aid in creating a more sustainable future. The increase of available measurements, growth of real time processing capacity and communication abilities is changing the opportunities in the power system landscape. There is a potential to utilize more measurement data, Reinforcement Learning and a FACTS device for optimization purposes.

The task would include: 
• Create suitable test networks
• Investigate and develop prototype algorithm based on latest advancements in deep Reinforcement Learning
• Train, tune and test the algorithms on the networks and demonstrate policy optimality

Requirements

We are looking for you who are studying a university master program within a relevant technical area along with an interest within artificial intelligence or Reinforcement Learning. Experience in Reinforcement Learning or programming is positive. We aim to start the thesis in early September.