https://dasya.itu.dk/semester/fall-2026/ Recent content in Fall 2026 on DASYA Hugo -- gohugo.io en-us © DASYA, IT University of Copenhagen Mon, 24 Aug 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/sbut-summary/ Mon, 24 Aug 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/sbut-summary/ A number of potential projects in the fields of extreme networking - long distance, e.g. space (satellites, stratospheric balloons) - LPWAN, LoRa, LoRaWan resilient and post-collapse networking - see also https://dasya.itu.dk/docs/pcn.html environmental sensors (from sea to stratosphere) Edge AI, embedded ML, TinyML IT and solar energy - see https://solar.itu.dk Please contact me for detail discussions. https://dasya.itu.dk/for-students/proposals/current/sbut-gpu-in-space/ Mon, 24 Aug 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/sbut-gpu-in-space/ Recently, large datacenters in space have been proposed. See Starcloud, Axiom, Thales Alenia, Google Suncatcher, SpaceX. Aside from the launch cost and many other challenges, one principal question is: How doe s GPU, producing significant amounts of heat, get rid of thermal energy? And even the solar panels, with efficiencies of around 30%, turining 70% of all incoming energy into thermal - how do they? This project should do the calculations as well as build a small demonstrator - a source of thermal energy, potentially a real GPU (we have various small Nvidia EdgeAI devices to burn …), to be deployed in a Mars simulator vacuum tank. https://dasya.itu.dk/for-students/proposals/current/zk-xli-dataprepare4qo/ Mon, 24 Aug 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/zk-xli-dataprepare4qo/ Traditional cost-based query optimizers are highly effective for many database workloads, especially for short and moderately complex queries. However, for queries involving many joins and large search spaces, optimization can become expensive and inaccurate. Recent learning-based query optimizers have shown promising results for complex queries, but they may introduce unnecessary overhead for simpler queries. This project aims to develop a hybrid query optimizer that combines the advantages of both approaches. The main idea is to analyze an incoming query and dynamically decide whether it should be handled by a traditional cost-based optimizer or a learning-based optimizer. https://dasya.itu.dk/for-students/proposals/current/pito-llm-agents/ Wed, 05 Aug 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/pito-llm-agents/ This project would investigate the storage access patterns of LLM agents using open-source state-of-the-art inference frameworks such as vLLM and SGLang and workload traces such as the ones released by Mooncake, Alibaba, Swiss AI Initiative, etc. Depending on the size and duration of the project, it can also propose or implement optimizations for the storage access. This project would be suitable as a standalone project, a research project with an MSc thesis followup, or BSc thesis. https://dasya.itu.dk/for-students/proposals/current/pito-dgxspark/ Wed, 05 Aug 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/pito-dgxspark/ This project investigates how running multiple deep learning tasks simultaneously (collocation) affects performance on resource-constrained edge devices, specifically the NVIDIA DGX Spark. Students will deploy and benchmark various model in isolated vs. different collocated scenarios, measure metrics such as GPU utilization, memory usage, training time, and analyze the trade-offs between throughput, fairness, and system efficiency. This project would be suitable as a BSc or MSc thesis. The workload can be adjusted based on the thesis and group size. https://dasya.itu.dk/for-students/proposals/current/pito-fdpsim-duckdb/ Wed, 05 Aug 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/pito-fdpsim-duckdb/ Traditionally solid-state drives (SSDs) does not give the users the ability to control the data placement on the SSD. This often leads to suboptimal performance and lowers SSD lifetime, since SSDs internally don’t allow in-place updates. The updated disk pages are written elsewhere and the old versions have to be garbage collected. This poses problems if data with different lifetimes and update frequencies often gets written in the same region. https://dasya.itu.dk/for-students/proposals/current/pito-gpu-db-perf/ Wed, 05 Aug 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/pito-gpu-db-perf/ This rise of hardware accelerators to meet the demand of AI workloads has also led to a variety of novel methods to leverage GPUs for traditional data analytics workloads. A key concern for any data-intensive system using GPUs is the efficiency of moving the data to the accelerator. In this project, we will investigate the performance of GPU databases and how they are impacted by the different kinds of data movement (CPU memory to GPU memory, SSD to GPU, etc. https://dasya.itu.dk/for-students/proposals/current/pito-llm-kvcache/ Wed, 05 Aug 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/pito-llm-kvcache/ This project would investigate key-value(KV)-cache offloading in large language model (LLM) inference engines, with particular emphasis on data placement and storage efficiency. First, the project would use an emulator such as FEMU to characterize workload access patterns, including hot-key distributions, write intensity, and write amplification factor (WAF). Then, we would evaluate the potential of Flexible Data Placement (FDP) for improving data placement decisions on SSDs for KV-cache accesses. The project will use open source state-of-the-art KV-cache types such as LMCache or the ones implemented by the inference frameworks such as vLLM and SGLang. https://dasya.itu.dk/for-students/proposals/archive/mh-zk-wayang-metadata-store/ Fri, 24 Jul 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/archive/mh-zk-wayang-metadata-store/ BSc / MSc Thesis. This project focuses on designing and implementing a metadata management store for Apache Wayang. Wayang lets users express a data processing task in a platform-agnostic way and then decides which underlying platform (e.g., Spark, Flink, Postgres, Java) should execute each part of the plan. Today, however, Wayang has no dedicated component for managing metadata: information such as schemas, statistics, or learned cost functions is handled implicitly by individual subsystems rather than stored and reused in one place. https://dasya.itu.dk/for-students/proposals/current/mh-ai-db-workloads/ Fri, 24 Jul 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/mh-ai-db-workloads/ BSc Thesis / MSc Thesis. The goal of this project is to use AI coding tools such as ChatGPT Codex and Claude Code together with optimization tools such as OpenEvolve to find innovative solutions for various database-y workloads, including: SIEM workloads, such as finding IP addresses in large security data sets String search workloads, such as finding substrings or evaluating regular expressions in large bodies of text Data analysis workloads, such as executing SQL queries over application logs The core idea is to find better statistics to summarize data partitions. https://dasya.itu.dk/for-students/proposals/current/mh-llm-rng-personas/ Fri, 24 Jul 2026 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/mh-llm-rng-personas/ BSc Thesis. This project builds on a previous master’s thesis, Prompting Luck: Bias and Manipulability in LLM-Generated Randomness, that explored replacing traditional random number generators (RNG) in games with large language models (LLMs). The idea is that an LLM can produce “intelligent” random numbers, for example the outcome of a slot machine spin, that take context into account rather than being drawn from a uniform distribution. The goal of this project is to experimentally validate the findings of that thesis by building an actual game, connecting it to an LLM, and putting the idea to the test with real players. https://dasya.itu.dk/for-students/proposals/current/zk-lrt/ Mon, 09 Jun 2025 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/zk-lrt/ Query optimization is crucial for any data management system to achieve good performance. Recent advancements in AI have led academia and industry to investigate learning-based techniques in query optimization. In particular, many works propose replacing the cost model used during plan enumeration with a machine learning model (typically a regression model) that estimates the runtime of a query plan. Interestingly though, it is well-known that what really matters in query optimization is the relative order of the query plan alternatives and not their actual cost or runtime. https://dasya.itu.dk/for-students/proposals/current/zk-apachewayang/ Wed, 28 May 2025 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/zk-apachewayang/ Are you interested in working with a big data open source project? You are welcome to conduct your thesis/project in Apache Wayang. Apache Wayang is the first cross-platform framework that allows users to specify their task/query in a system-agnostic manner and Wayang will determine which is the best system(s) to execute this task with the goal of optimizing performance. For a general overview check this paper. Potential Intended Learning Outcomes depending on the concrete topic: https://dasya.itu.dk/for-students/proposals/current/zk-kges/ Wed, 28 May 2025 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/zk-kges/ Knowledge graphs (KGs) are extensively used in many application domains, such as search engines, product recommendation, and bioinformatics. Knowledge graph completion (a.k.a.~link prediction), i.e.,~the task of inferring missing information from knowledge graphs, is a widely used task in the above applications. This project will investigate how to loosely-couple the data-driven power of knowledge graph embeddings with domain-specific reasoning stemming from experts or reasoning systems. In this way, we not only enhance the prediction accuracy with domain knowledge that may not be included in the input knowledge graph but also allow users to plugin their own knowledge graph embedding and reasoning method. https://dasya.itu.dk/for-students/proposals/current/zk-apachewayang-mcp/ Wed, 28 May 2025 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/zk-apachewayang-mcp/ Are you interested in working with a big data open source project and AI? You are welcome to conduct your thesis/project in the context of Apache Wayang. Apache Wayang is the first cross-platform framework that allows users to specify their task/query in a system-agnostic manner and Wayang will determine which is the best system(s) to execute this task with the goal of optimizing performance. For a general overview check this paper. https://dasya.itu.dk/for-students/proposals/current/zk-apachewayang-green/ Wed, 28 May 2025 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/zk-apachewayang-green/ Are you interested in working with a big data open source project and help the environment? You are welcome to conduct your thesis/project in Apache Wayang. Apache Wayang is the first cross-platform framework that allows users to specify their task/query in a system-agnostic manner and Wayang will determine which is the best system(s) to execute this task with the goal of optimizing performance. For a general overview check this paper. Wayang currently optimises query execution time by selecting the right platforms. https://dasya.itu.dk/for-students/proposals/current/zk-datafarm/ Wed, 28 May 2025 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/zk-datafarm/ Query optimization is crucial for any data management system to achieve good performance. Recent advancements in AI have led academia and industry to investigate learning-based techniques in query optimization. In particular, many works propose replacing the cost model used during plan enumeration with a machine learning model that estimates the runtime of a plan. However, to build such a model lots of training data are required. In this context, training data comprise of query plans and their runtimes. https://dasya.itu.dk/for-students/proposals/archive/osk-ecp-bench/ Tue, 14 Jan 2025 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/archive/osk-ecp-bench/ The extended Cluster Pruning (eCP) index, is a hierarchical approximate nearest neighbor index. eCP-FS is a “white-box” implementation of the index as a file structure using zarr. The index building is done through Python, while the index loading and search have been implemented in both pure Python and a Rust-Python package. While the index is slower from disk than other disk-based ANN solutions due to it not being stored as a single or few binary files but tons of smaller files, it allows more control over what is loaded, is easy to extend and when in-memory is on par with other memory based indexes. https://dasya.itu.dk/for-students/proposals/current/vs-similarity/ Fri, 22 Nov 2024 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/vs-similarity/ Machine learning methods are often evaluated on benchmark datasets, in computer vision, medical imaging, NLP and other fields. In such evaluation, researchers often describe the data as being: representative, for example based on the distribution of ages of the patients mirroring the world population, similar, for example because both dataset contain pictures of animals diverse, for example because data was collected from multiple countries Recent research shows, the concepts of representativity, similarity, and diversity, are not strictly defined, and different notions might be used by different researchers. https://dasya.itu.dk/for-students/proposals/current/jucp-cubesat-lora/ Mon, 01 Jul 2024 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/jucp-cubesat-lora/ LoRa is a long range, low bandwith networking protocol widely used in Internet of Things projects, sensor networks, low power, low cost and embedded systems. LoRa’s encoding schema allows for extremely long distance communications with small power usage and small simple antennas. This combination of features has made it attractive to small satellite operators flying cubesats and LoRa is now routinely used to connect small ground based nodes to satellites, typically for remote sensor network deployments, and for satellite telemetry, e. https://dasya.itu.dk/for-students/proposals/current/danishstudentcubesat/ Thu, 09 Sep 2021 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/danishstudentcubesat/ The Danish Student Cubesat Program is an inter university collaboration that will launch 3 cubesats into Low Earth Orbit over the next 4 years. The satellites will be designed, operated, programmed and built by students and the project offers an opportunity for Master’s students to take part in a live satellite project. ITU is partnering with Aarhus University on DISCOSAT2 which will be an imaging instrument built to photograph and thermally image the Arctic ocean and Greenland in collaboration with the Aarhus Arctic Research Centre. https://dasya.itu.dk/for-students/proposals/current/cities1/ Thu, 01 Jul 2021 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/cities1/ The idea behind “15-minutes cities” is that within a short walk or bike ride people should have access to all necessary facilities that constitute the essence of urban living, such as parks, shops, cafes, schools, hospitals. Initiatives to transform cities according to this paradigm are currently being implemented across the world, in an attempt to make urban spaces more liveable, sustainable and community-oriented. However, a systematic understanding of the connection between urban form and accessibility is missing. https://dasya.itu.dk/for-students/proposals/current/cycling/ Thu, 01 Jul 2021 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/cycling/ As a response to increased traffic congestion and the need to reduce carbon emissions, cities consider ways to modernise, build and extend transit systems. Transit network design solutions can benefit from analysing the large amount of crowd-sourced location data available, which provides valuable insights into population mobility needs. Designing efficient metro lines, bicycle paths, or bus routes brings a number of conflicting constraints into play: from the user’s perspective, an efficient transit network is easily accessible and time-efficient, while from the provider’s side there is, amongst others, a limitation on the budget. https://dasya.itu.dk/for-students/proposals/current/cities2/ Thu, 01 Jul 2021 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/cities2/ The idea behind “15-minutes cities” is that within a short walk or bike ride people should have access to all necessary facilities that constitute the essence of urban living, such as parks, shops, cafes, schools, hospitals. Initiatives to transform cities according to this paradigm are currently being implemented across the world, in an attempt to make urban spaces more liveable, sustainable and community-oriented. However, a systematic understanding of the connection between urban form and accessibility is missing. https://dasya.itu.dk/for-students/proposals/current/music_embeddings/ Thu, 01 Jul 2021 00:00:00 +0000 https://dasya.itu.dk/for-students/proposals/current/music_embeddings/ Musical genres are inherently ambiguous and difficult to define. Even more so is the task of establishing how genres relate to one another. Yet, genre is perhaps the most common and effective way of describing musical experience. The number of possible genre classifications (e.g. Spotify has over 4000 genre tags, LastFM over 500,000 tags) has made the idea of manually creating music taxonomies obsolete. The aim of this project is to create hyperbolic embeddings to learn a general music taxonomy based on the co-occurrence of genres.