5 papers
Formal Disco: Scalable Open-Ended Generation of Formally Verified Programs
Gabriel Poesia, Simon Henniger, Tzu-Han Hsu +2
The cost of producing code is rapidly diminishing with increasingly capable AI agents, while quality assurance of generated programs has not kept pace. Formal verification provides…
HyPOLE: Hyperproperty-Guided Multi-Agent Reinforcement Learning under Partial Observation
Arshia Rafieioskouei, Tzu-Han Hsu, Matthew Lucas +1
Formal specification is a powerful tool to guide the learning process and provides significant advantages over reward shaping: (1) mathematical rigor; (2) expressiveness to specify…
HypRL: Reinforcement Learning of Control Policies for Hyperproperties
Tzu-Han Hsu, Arshia Rafieioskouei, Borzoo Bonakdarpour
Reward shaping in multi-agent reinforcement learning (MARL) for complex tasks remains a significant challenge. Existing approaches often fail to find optimal solutions or cannot ef…
HyperQB: A Bounded Model Checker for Hyperproperties
Tzu-Han Hsu, Milad Rabizadeh, Kenneth Rogale +4
We introduce the tool HyperQB 2.0, the first highly efficient push-button bounded model checker (BMC) for hyperproperties. HyperQB takes as input a model in NuSMV or Verilog and a…
Syntax-Guided Automated Program Repair for Hyperproperties
Raven Beutner, Tzu-Han Hsu, Borzoo Bonakdarpour +1
We study the problem of automatically repairing infinite-state software programs w.r.t. temporal hyperproperties. As a first step, we present a repair approach for the temporal log…