14 citations · 33 across the 22 of their papers we have counts for
23 papers
Learning the Signature of Memorization in Autoregressive Language Models
David Ilić, Kostadin Cvejoski, David Stanojević +1
All prior membership inference attacks for fine-tuned language models use hand-crafted heuristics (e.g., loss thresholding, Min-K\%, reference calibration), each bounded by the des…
Protecting Private Code in IDE Autocomplete using Differential Privacy
Evgeny Grigorenko, David Stanojević, David Ilić +2
Modern Integrated Development Environments (IDEs) increasingly leverage Large Language Models (LLMs) to provide advanced features like code autocomplete. While powerful, training t…
Control Models for In-IDE Code Completion
Aral de Moor, Yana Hrynevich, Hleb Badzeika +6
We introduce control models for LLM-powered code completion in JetBrains IDEs: ML classifiers which trigger inference and filter the generated suggestions to better align them with…
Domain-Adaptation through Synthetic Data: Fine-Tuning Large Language Models for German Law
Ali Hamza Bashir, Muhammad Rehan Khalid, Kostadin Cvejoski +7
Large language models (LLMs) often struggle in specialized domains such as legal reasoning due to limited expert knowledge, resulting in factually incorrect outputs or hallucinatio…
Powerful Training-Free Membership Inference Against Autoregressive Language Models
David Ilić, David Stanojević, Kostadin Cvejoski
Fine-tuned language models pose significant privacy risks, as they may memorize and expose sensitive information from their training data. Membership inference attacks (MIAs) provi…
On Foundation Models for Temporal Point Processes to Accelerate Scientific Discovery
David Berghaus, Patrick Seifner, Kostadin Cvejoski +1
Many scientific fields, from medicine to seismology, rely on analyzing sequences of events over time to understand complex systems. Traditionally, machine learning models must be b…
Towards Fast Coarse-graining and Equation Discovery with Foundation Inference Models
Manuel Hinz, Maximilian Mauel, Patrick Seifner +3
High-dimensional recordings of dynamical processes are often characterized by a much smaller set of effective variables, evolving on low-dimensional manifolds. Identifying these la…
In-Context Learning of Temporal Point Processes with Foundation Inference Models
David Berghaus, Patrick Seifner, Kostadin Cvejoski +2
Modeling event sequences of multiple event types with marked temporal point processes (MTPPs) provides a principled way to uncover governing dynamical rules and predict future even…
Multi-Modal Vision vs. Text-Based Parsing: Benchmarking LLM Strategies for Invoice Processing
David Berghaus, Armin Berger, Lars Hillebrand +2
This paper benchmarks eight multi-modal large language models from three families (GPT-5, Gemini 2.5, and open-source Gemma 3) on three diverse openly available invoice document da…
In-Context Learning of Stochastic Differential Equations with Foundation Inference Models
Patrick Seifner, Kostadin Cvejoski, David Berghaus +2
Stochastic differential equations (SDEs) describe dynamical systems where deterministic flows, governed by a drift function, are superimposed with random fluctuations, dictated by…
Foundation Inference Models for Markov Jump Processes
David Berghaus, Kostadin Cvejoski, Patrick Seifner +2
Markov jump processes are continuous-time stochastic processes which describe dynamical systems evolving in discrete state spaces. These processes find wide application in the natu…
Zero-shot Imputation with Foundation Inference Models for Dynamical Systems
Patrick Seifner, Kostadin Cvejoski, Antonia Körner +1
Dynamical systems governed by ordinary differential equations (ODEs) serve as models for a vast number of natural and social phenomena. In this work, we offer a fresh perspective o…
Neural Dynamic Focused Topic Model
Kostadin Cvejoski, Ramsés J. Sánchez, César Ojeda
Topic models and all their variants analyse text by learning meaningful representations through word co-occurrences. As pointed out by Williamson et al. (2010), such models implici…
The future is different: Large pre-trained language models fail in prediction tasks
Kostadin Cvejoski, Ramsés J. Sánchez, César Ojeda
Large pre-trained language models (LPLM) have shown spectacular success when fine-tuned on downstream supervised tasks. Yet, it is known that their performance can drastically drop…
Hidden Schema Networks
Ramsés J. Sánchez, Lukas Conrads, Pascal Welke +2
Large, pretrained language models infer powerful representations that encode rich semantic and syntactic content, albeit implicitly. In this work we introduce a novel neural langua…
Informed Pre-Training on Prior Knowledge
Laura von Rueden, Sebastian Houben, Kostadin Cvejoski +2
When training data is scarce, the incorporation of additional prior knowledge can assist the learning process. While it is common to initialize neural networks with weights that ha…
Knowledge Augmented Machine Learning with Applications in Autonomous Driving: A Survey
Julian Wörmann, Daniel Bogdoll, Christian Brunner +49
The availability of representative datasets is an essential prerequisite for many successful artificial intelligence and machine learning models. However, in real life applications…
Combining expert knowledge and neural networks to model environmental stresses in agriculture
Kostadin Cvejoski, Jannis Schuecker, Anne-Katrin Mahlein +1
In this work we combine representation learning capabilities of neural network with agricultural knowledge from experts to model environmental heat and drought stresses. We first d…
Dynamic Review-based Recommenders
Kostadin Cvejoski, Ramses J. Sanchez, Christian Bauckhage +1
Just as user preferences change with time, item reviews also reflect those same preference changes. In a nutshell, if one is to sequentially incorporate review content knowledge in…
Generative Deep Learning Techniques for Password Generation
David Biesner, Kostadin Cvejoski, Bogdan Georgiev +2
Password guessing approaches via deep learning have recently been investigated with significant breakthroughs in their ability to generate novel, realistic password candidates. In…
Recurrent Point Review Models
Kostadin Cvejoski, Ramses J. Sanchez, Bogdan Georgiev +2
Deep neural network models represent the state-of-the-art methodologies for natural language processing. Here we build on top of these methodologies to incorporate temporal informa…
Recurrent Point Processes for Dynamic Review Models
Kostadin Cvejoski, Ramses J. Sanchez, Bogdan Georgiev +3
Recent progress in recommender system research has shown the importance of including temporal representations to improve interpretability and performance. Here, we incorporate temp…
Adiabatic Quantum Computing for Binary Clustering
Christian Bauckhage, Eduardo Brito, Kostadin Cvejoski +3
Quantum computing for machine learning attracts increasing attention and recent technological developments suggest that especially adiabatic quantum computing may soon be of practi…