Institution profile

University of Zagreb

Academic institutioneurope · hr
Official website
Research library142linked papers
Opportunities0open roles
Selected work

Representative Papers

Reinforcement Learning with Stochastic Reward Machines

Jun 28, 2022AAAI Conference on Artificial Intelligence

Existing reward machines (RMs) assume noise-free reward signals, limiting their applicability in real-world settings characterized by sparse, action-sequence-dependent, and noisy rewards. Method: This paper proposes the Stochastic Reward Machine (SRM), a novel RM variant that explicitly models stochastic reward observations. We introduce constraint solving into RM learning for the first time, enabling automatic inference of state partitions and transition relations from agent exploration trajectories to synthesize a minimal SRM. Theoretical analysis establishes asymptotic convergence to an optimal policy under reward noise. Results: Experiments on two representative noisy-reward tasks demonstrate that our approach significantly outperforms existing RM-based methods and naive denoising baselines, validating its robustness and effectiveness in learning from unreliable reward signals.

30 citations1 influentialRead paper

Dynamic Software Updating in Java: Comparing Concepts and Resource Demands

Apr 03, 2017International Conference on the Art, Science and Engineering of Programming

Dynamic Software Updates (DSUs) reduce downtime, improve security, and accelerate iterative development; however, existing research lacks systematic, empirical comparisons of DSU techniques in the Java ecosystem—particularly regarding supported code-change types and associated runtime resource overhead. Method: This paper presents the first cross-cutting evaluation of mainstream Java DSU mechanisms—including bytecode hot-swapping, class redefinition, and proxy injection—using micro-benchmarks and system-level monitoring to quantify CPU and memory overhead across representative change scenarios (e.g., method body modification, field addition/removal). Contribution/Results: We propose a “change capability–resource cost” analytical framework that empirically correlates update expressiveness with runtime overhead, filling a critical gap in evidence-based DSU comparison. Our findings provide data-driven guidance for industry practitioners selecting lightweight, secure, and evolution-aware hot-update solutions aligned with real-world maintenance requirements.

3 citations1 influentialRead paper

Old Habits Die Hard: How Conversational History Geometrically Traps LLMs

Feb 08, 2026

This study investigates behavioral biases in large language models induced by dialogue history, with a particular focus on how hallucinations from prior interactions persistently distort subsequent responses. To address this, the authors propose the History-Echoes framework, which for the first time models history-induced generation bias as “geometric traps” in latent space, revealing a strong correlation between probabilistic state consistency and geometric consistency of hidden representations. Methodologically, the approach combines Markov chain modeling of dialogue states with a geometric consistency metric derived from continuous hidden representations to quantify historical bias. Extensive experiments across three prominent model families and six datasets validate the high correlation between these dual perspectives, demonstrating that dialogue history can indeed form geometric traps in latent space that constrain the model’s generation trajectory.

1 citationsRead paper
Recent publications

Latest Papers