Here you can find information about the projects that you can join – what you have heard at this year’s Project Forum, and some others. The list will be updated with incoming information throughout the school year.
Projektek:
- Expert systems to support medical and biochemical research using artificial intelligence, data mining and web mining
Goals, Topics: The project carries out research activities that automate the processing of thousands of scientific results (articles) related to specific topics using web mining, data mining, text mining technologies, and build mathematical and IT models of hidden information for specific health purposes.
Used Technologies: Artificial intelligence, data mining and web mining.
Who are we searching for?: Students interested in Artificial Intelligence are welcome.
How much time a student needs to contribute: 8-20 hours weekly.
Contact person: Kocsisné Dr. Szilágyi Gyöngyi
Contact Information: szilagyi at aszt.inf.elte.hu - Static code analysis
Goals, Topics: Static analysis of source code to detect possible errors and to help code comprehension. Our tools are public domain and used daily by companies like Google, Apple, Sony, Ericsson, Cloudera, etc. Visit https://github.com/ericsson/codechecker and https://github.com/ericsson/codecompass
Used Technologies: LLVM/Clang, Clang Tidy, Clang Static Analyzer, SQL, Machine Learning
Who are we searching for?: Students with strong programming competence in C++ and/or Java, Python, Golang, Javascript. Basic knowledge on compiler technology is an advantage but not necessary. There are various tasks in the projects from hardcore C++ to implementing GUI in Javascript, so everybody who is interested is welcome.
How much time a student needs to contribute: 10-20h weekly
Contact person: Porkoláb Zoltán, Gera Zoltán
Contact Information: gsd at inf.elte.hu, gerazo at caesar.elte.hu - Edge Computing / 5G
Goals, Topics: The project is in constant search after use cases for edge computing. We implement real scenarios, deploy them on real equipment and investigate usability and pitfalls. Our current direction is towards 3D computer vision, high bandwidth, low delay streaming, cloud technologies, ML and measurements.
Used Technologies: OpenCV, GStreamer, ElastisSearch, Kibana, C++, Python, git, Tensorflow, RPis and routers
Who are we searching for?: Anyone who is motivated enough should come. This project uses a lot of different techs. We will help you get a grasp on them.
How much time a student needs to contribute: Weekly 1-2 days is beneficial.
Contact person: Zoltán Gera
Contact Information: gerazo at caesar.elte.hu - Human-Machine Co-Learning on Echocardiograms
Goals, Topics: The project belongs to the Department of Artificial Intelligence as a cooperative research with Semmelweis University. Goals in keywords: feature extraction, language development.
Used Technologies: deep learning
Who are we searching for?: MSc students with excellent programming skills in Python, PyTorch, TensorFlow.
How much time a student needs to contribute: 20 hours weekly
Contact person: Belics Éva
Contact Information: belicseva at inf.elte.hu - PTSD Data Clustering
Goals, Topics: The project belongs to the Department of Artificial Intelligence as a cooperative research with Rush University Medical Center. Goals in keywords: clustering.
Used Technologies: deep learning
Who are we searching for?: MSc students with excellent programming skills in Python, PyTorch, TensorFlow.
How much time a student needs to contribute: 20 hours weekly
Contact person: Belics Éva
Contact Information: belicseva at inf.elte.hu - Evaluation in Group Formation Task
Goals, Topics: The project belongs to the Department of Artificial Intelligence as a cooperative research with Carnegie Mellon University. Goals in keywords: analysis of face, body and speech data.
Used Technologies: deep learning
Who are we searching for?: MSc students with excellent programming skills in Python, PyTorch, TensorFlow.
How much time a student needs to contribute: 20 hours weekly
Contact person: Belics Éva
Contact Information: belicseva at inf.elte.hu - TensorBoard-NipgBoard Development
Goals, Topics: The project belongs to the Department of Artificial Intelligence as a cooperative research with Argus Cognitive. Goals in keywords: frontend and backend development.
Used Technologies: deep learning
Who are we searching for?: MSc students with excellent programming skills in Python, PyTorch, TensorFlow.
How much time a student needs to contribute: 20 hours weekly
Contact person: Belics Éva
Contact Information: belicseva at inf.elte.hu - Speed Estimation without Semantic Information
Goals, Topics: The project belongs to the Department of Artificial Intelligence as a cooperative research with BOSCH. Goals in keywords: integration of computer vision and machine vision.
Used Technologies: deep learning
Who are we searching for?: MSc students with excellent programming skills in Python, PyTorch, TensorFlow.
How much time a student needs to contribute: 20 hours weekly
Contact person: Belics Éva
Contact Information: belicseva at inf.elte.hu - Melanoma Detection
Goals, Topics: The project belongs to the Department of Artificial Intelligence as a cooperative research with Degetel and Semmelweis University. Goals in keywords: skin lesion registration and change detection on the back.
Used Technologies: deep learning
Who are we searching for?: MSc students with excellent programming skills in Python, PyTorch, TensorFlow.
How much time a student needs to contribute: 20 hours weekly
Contact person: Belics Éva
Contact Information: belicseva at inf.elte.hu - Machine Learning for Static Analysis
Goals, Topics: Clang compiler’s built-in static analyzer finds lot of potential bugs. This research tries to find ways of predicting whether an issue will be real or not, serious or not.
Used Technologies: clang compiler, clang static analyzer, C/C++, Python, XGBoost as ML library (gradient boosting), Git, Linux scripting
Who are we searching for?: We are searching for motivated people interested in either C++, machine learning or computer language analysis.
How much time a student needs to contribute: 0.5-2 days per week
Contact person: Gera Zoltán
Contact Information: gerazo at caesar.elte.hu - Static Analysis for Security
Goals, Topics: In this project, we are searching for potential security problems that programmers usually make and improve Clang’s bug detection suite so it will recognize these situations and help programmers in avoiding such failures.
Used Technologies: clang-tidy, clang static analyzer, C/C++, git
Who are we searching for?: If you are interested in getting deeper knowledge of C/C++ or you are moved by making the world more secure, this is the project for you!
How much time a student needs to contribute: 0.5-1.5 days per week
Contact person: Gera Zoltán
Contact Information: gerazo at caesar.elte.hu - Development and Testing at Scale
Goals, Topics: Extending a compiler and development environment, to help the development of large scale test systems. We also aim to gain knowledge on large scale production and test systems.
Used Technologies: Java, Eclipse, ANTLR; and we analyze TTCN-3 projects.
Who are we searching for?: Students looking for thesis topics and students who would like to discover the properties of large scale software systems. Some Java knowledge is required.
How much time a student needs to contribute: For scholarship 10 hours/week. For credits a few hours/week.
Contact person: Kovács Attila
Contact Information: Déli Tömb 2-207 - Mediso Medical Imaging Research
Goals, Topics: General goals is to enhance image quality with various reconstruction and analysis methods to fasten up and automate clinical and preclinical protocols
Used Technologies: Python(Numpy, Simpy), C++(CMake, Qt)
Who are we searching for?: We are looking for creative and talented programmers whos interested in Medical Imaging and Medical Image Analysis. We have different trendy topics such as cancer research, cardiac diagnostic research. We mostly looking for candidates with strong problem solving and coding skills.
How much time a student needs to contribute: 10 hours per week
Contact person: Szűcs Ádám
Contact Information: 0123pv3210 at gmail.com - P4 project
Goals, Topics: Applicability of the P4 language (implementation of interesting telecommunication case studies motivated by industrial partners). Extension possibilities of the language. Implementing an effective compiler. Verification and (static) analysis of P4 programs.
Used Technologies: Python, C, P4, DPDK
Who are we searching for?: basic knowledge of telecommunication networks or compilers
How much time a student needs to contribute: 1 day per week
Contact person: Tejfel Máté
Contact Information: matej at inf.elte.hu - IoT&Data Innovation Lab
Goals, Topics: IoT&Data Innovation Lab with around 20 industrial partners, 14 ongoing R&D&I projects is one of the largest research group within ELTE Faculty of Informatics. The Lab was founded in 2016, equipped with a variety of devices (including sensors, cameras, controllers, 3D printer, robots, a knitting machine, and the latest IBM Power9 AI server. We have a small physical lab within EIT Digital – a proof-of-concept ready living lab environment to develop and validate. IoT&Data Innovation Lab focuses on: AgriTech, FoodTech, eHealth-MedTech and FinTech projects. We offer technical (IoT, Data Analytics and AI, development) and business support, plus Incubation service for students, professors, researchers and industrial partners. After a slow and difficult start the lab modell seems to be working, we can offer scholarships.
Working team: 21 person (5 professors-researchers, 7 external domain experts, 9 students).
The projects you can join to:
MedTech/eHealth
– Supporting a movement therapy development for children: using Kinect / camera, sensors, etc.
– Part of the largest Hungarian medical data analytics project we are building a Medical Datalake: Our task for the next months is to build up a team with the knowledge of data analytics + AI and to test medical data analytics frameworks and datalake solutions.
AgriTech and FoodTech field:
– Vertical farming: our smart greenhouse prototype is ready, the next step is to install it to a real environment and to collect and analyse data.
– Rabbit breeding: we have 3 sub-projects: 1. Pland and desing a smart cage. 2. Analyse the breeding process, plan and develop an information system with data analytics services for our international industrial partner. 3. Predict foot disease based on image processing.
– Analysing of food industry processes: we have two industrial partners within the food processing industry. Sub-projects: 1. identify and spot defected products in real time (on the conveyor belt) using cameras. 2. One of our industrial partner is organizing site visits to some potential food processing fanctories. Our main task theree is to analyse business processes and identify digitalization points in order to define projects.
– Bee keeping: Develop a bee counter tool.
Thesis / TDK works:
– Develop a farm-to-fork blockchain service for monitoring luxury products
– Analyse agribusiness datasets in Python / PowerBI
Logistics and Smart City
– Occupancy display for trains by big data analysis
– Real-time timetable prediction for trains
– Develop a air pollution meter for schools and for bikers using Microsoft Azure and SAP technologies.
Used Technologies: Depends on the project
Who are we searching for?: Python / ML-DL / Arduino / Image Processing
How much time a student needs to contribute: 4-20 hours weekly
Contact person: Tarcsi Ádám
Contact Information: ade at inf.elte.hu - RefactorErl
Goals, Topics: Static/dynamic source code analysis and transformations for Erlang.
Used Technologies: –
Who are we searching for?: Somebody who is willing to learn new technologies and tools and want to work on an interesting R&D tasks.
How much time a student needs to contribute: 4 hours/week is a minimum in a Lab like RefactorErl, but if you want to achieve nice result, you should find a bit more time for the project.
Contact person: Tóth Melinda, Bozó István
Contact Information: tothmelinda at elte.hu, bozo_i at inf.elte.hu - High-Assurance Refactoring Project (HARP)
Goals, Topics: This project develops a methodology for verifying architectures of refactoring tools, and demonstrate that methodology in practice by mechanically verifying, using the Coq theorem prover, a scheme-based refactoring system for programs in the Erlang language.
We also aim for demonstrating how the results can be transferred to provide assurance for existing Erlang refactoring tools, including RefactorErl and Wrangler. Although the demonstrations will be for a single programming language, the approach developed by the project should be transferable to other languages and tools.
Used Technologies: formal semantics, type theory, logic, language design, Coq, Erlang, Java
Who are we searching for?: We can provide both theoretical and practical tasks, or the combination of the two. At the very least we require you to have some interest and/or experience in refactorings and/or programming theory and/or formal semantics and/or language&compiler design&implementation.
How much time a student needs to contribute: From 8 to 20 hours a week, including a 2-hour-long meeting, based on personal goals and the chosen task.
Contact person: Németh Dávid J. , Horpácsi Dániel
Contact Information: ndj at inf.elte.hu, daniel-h at elte.hu - Task Oriented Programming
Goals, Topics: Task-Oriented Programming (TOP) is a new programming paradigm for developing interactive multi-user systems. This programming paradigm uses “Tasks” as the central concept.
The main motivation behind the paradigm is to support programmers in solving complex problems on a high abstraction level. If we can precisely define a smaller task or can decompose a complex one into smaller simpler tasks, we should have enough information to generate executable systems by using only the task specifications. This means that all technical infrastructure of an interactive system (handling events, storing data, encoding/decoding, etc.) can be factored out and can be solved generically.
The paradigm is not only scientifically interesting, but it also has industrial applications. The main users of the ToP framework are the Netherlands Coast Guard, the Netherlands Organisation for Applied Scientific Research and the Dutch Defense Academy (NLDA).
Used Technologies: Clean
Who are we searching for? Students well-versed in functional programming. We have many different subprojects to choose from suited for any skill level (bachelor, masters and even PhD).
How much time a student needs to contribute: A few hours a week
Contact person: Péter Podlovics
Contact Information: anabra at inf dot elte dot hu - Whole program compilation for functional languages
Goals, Topics: Today, most compilers compile each translation unit individually, which means that each TU is analyzed and optimized independently. In contrast to this, whole program compilers, see all the TUs together as a big program. This means the compiler has much more information to work with, thus it can generate much more efficient code. At first, we may think that whole program compilation might scale poorly. However, many popular industrial-strength compilers use whole program compilation, such as LLVM from the imperative world and MLton from the functional world.
On another note, functional languages might have very sophisticated compiler front ends, but they usually lack behind in terms of their back ends. This is mainly because most research around functional languages target the compilers’ front ends (e.g.: the type system). One of our main goals is to solve this problem by applying whole program compilation, and through dataflow analyses and aggressive code optimizations, generate performant machine code from high-level functional programs.
Used Technologies: Haskell, Datalog (for defining scalable dataflow analyses)
Who are we searching for? Students well-versed in functional languages, and interested in compilers.
How much time a student needs to contribute: A few hours a week
Contact person: Péter Podlovics
Contact Information: anabra at inf dot elte dot hu - Geoinformatics and Remote Sensing
Goals, Topics: Current topics are focused on object recognition, change detection and model reconstruction based on 3D LiDAR (Light Detection and Ranging) point clouds. The topics cover a wider range:
– Change detection of vegetation from airborne LiDAR Data
– Building recognition and change detection from airborne LiDAR data
– Object recognition and error analysis of railway infrastructure from mobile laser mapping
– Autonomous mapping of LiDAR point clouds based on sensor fusion and SLAM algorithms
Used Technologies: LiDAR, point clouds, digital elevation models, PCL, PDAL, GDAL
Who are we searching for?: MSc or PhD students with programming competence mainly in C++ and/or C#. Specialized knowledge on LiDAR technologies are not mandatory beforehand, but completing the relevant courses in the MSc programme is a benefit.
How much time a student needs to contribute: 10-16h weekly
Contact person: Máté Cserép
Contact Information: mcserep at inf.elte.hu - Task Management System
Goals, Topics: The aim of the project is the development of the Task Management System (TMS), which is a – soon to be – open source web-based assignment and examination management system for universities, primarily focused on programming and computer science related courses. The main features of the application are the integrated evaluation, testing, version control and plagiarism check of the student submissions.
Used Technologies: web services, REST, MVC, PHP, Yii, Docker
Who are we searching for?: Students with knowledge, experience (or at least interest) around the web technologies, the MVC design pattern and at least basic competence in PHP programming. Skills in other programming languages (e.g. C#, Java, Swift) are a benefit.
How much time a student needs to contribute: 8-12h weekly
Contact person: Máté Cserép
Contact Information: mcserep at inf.elte.hu
Summary information – Project Forum 2019.