Professional homepage of Thomas Ferrère
Senior Hardware Engineer, Imagination Technologies.
Research
- Datapath design
- Computer arithmetic
- Runtime verification
- Embedded systems
News
Short CV
I am a senior hardware engineer at Imagination Technologies, Kings Langley, UK.
Previously, I was postdoctoral researcher at IST Austria, Klosterneuburg, Austria in the group of Thomas A. Henzinger.
I completed my PhD at Verimag, University of Grenoble, France under the supervision of Oded Maler and in collaboration with Mentor Graphics corporation.
Previous Events
PhD Thesis
Assertions and Measurements for Mixed-Signal Simulation, University of Grenoble, 2016.
Publications
Preprints by topic. See also DBLP and Google Scholar.
Copyright belongs to respective publishers.
- Monitorability
We defined automata models capturing the (online) monitorability of discrete languages beyond finite-state.
We invented a strongly convergent algorithm to monitor the frequency of Markov-distributed events.
- Real-time Logics
We extended a result on the decidability of real-time temporal logic, based on infinitely precise semantics and finitely precise syntax.
We studied the (offline) monitoring of temporal logic extended with clock variables or mixed discrete-/continuous-time modalities.
We proposed an algorithm to monitor a first-order logic with real-valued variables, in which properties such as eventual stability can be described.
- Robust Monitoring
We invented an efficient algorithm to compute the continuous satisfaction degree (robustness) of a temporal formula over real-valued signals.
We proposed a variant of this algorithm to perform parametric identification.
We redefined the continuous satisfaction degree according to a system's interface and showed that resulting notions of vacuity and robustness simplify the falsification problem.
- Fault Localization
We demonstrated how segments of a signal falsifying a given formula can be identified.
We combined this analysis with formal and statistical techniques to accelerate the localisation of faults in a white-box system.
- Pattern Matching
We invented an offline algorithm to match regular expressions over continuous signals.
We adapted this algorithm to operate online through the notions of regular expression derivatives and timed automata.
We associated regular expressions over real-valued signals with an (offline-computable) contiuous satisfaction degree.
- Pattern Measuring
We demonstrated how to use regular patterns to specify and extract measurements over real-valued signals: firstly, by associating each matched segment with an aggregation function such as average or min/max; secondly, by decomposing the matched segment into a series of basic shapes through statistical regression.
- Tools
The above monitoring, pattern matching, measurement and diagnostics algorithms are implemented in an academic software tool.