Search

To increase the scientific output of particle physics experiments, upgrades are underway at all major accelerator facilities to significantly improve the luminosity. Consequently, the solid-state detectors used in the experiments will exhibit more severe radiation-induced damage. To ensure sufficiently long sensor lifetimes, alternative materials to the established silicon sensors, with improved resilience to radiation, are investigated. For one of the promising candidate materials, silicon carbide, only recently a model describing the radiation damage in technology aided computer design (TCAD) simulations has been proposed. In this paper we present our latest achievements towards modeling radiation damage of 4H-SiC in TCAD tools. We first verify the utilized TCAD framework against published silicon data and then use it to approximate measurements of neutron-irradiated 4H-SiC particle detectors. We are able to confirm in simulations the measurement results, i.e., an almost flat capacitance as a function of bias voltage and a decreasing forward current with increasing particle fluence. Based on our simulations we are able to explain the latter by trapped charge carriers that create a space charge region within the device., Comment: 4 pages, 4 figures

In the classical Approximate Majority problem with two opinions there are agents with Opinion 1 and with Opinion 2. The goal is to reach consensus and to agree on the majority opinion if the bias is sufficiently large. It is well known that the problem can be solved efficiently using the Undecided State Dynamics (USD) where an agent interacting with an agent of the opposite opinion becomes undecided. In this paper, we consider a variant of the USD with a preferred Opinion 1. That is, agents with Opinion 1 behave stubbornly -- they preserve their opinion with probability $p$ whenever they interact with an agent having Opinion 2. Our main result shows a phase transition around the stubbornness parameter $p \approx 1-x_1/x_2$. If $x_1 = \Theta(n)$ and $p \geq 1-x_1/x_2 + o(1)$, then all agents agree on Opinion 1 after $O(n\cdot \log n)$ interactions. On the other hand, for $p \leq 1-x_1/x_2 - o(1)$, all agents agree on Opinion 2, again after $O(n\cdot \log n)$ interactions. Finally, if $p \approx 1-x_1/x_2$, then all agents do agree on one opinion after $O(n\cdot \log^2 n)$ interactions, but either of the two opinions can survive. All our results hold with high probability.

The games research community has developed substantial knowledge on designing engaging experiences that draw players in. Surprisingly, less is known about player \textit{dis}engagement, with existing work predominantly addressing disengagement from the perspective of problematic play, and research exploring player disengagement from a constructive designer perspective is lacking. In this paper, we address this gap and argue that disengagement from games should be constructively designed, allowing players to exit play sessions in a self-determined way. Following a two-phase research approach that combines an interview study (n=16) with a follow-up online survey (n=111), we systematically analyze player perspectives on exiting play sessions. Our work expands the existing notion of disengagement through a characterization of exit experiences, a lens on disengagement as a process, and points for reflection for the design of games that seek to address player disengagement in a constructive way.

Logical fallacies are common errors in reasoning that undermine the logic of an argument. Automatically detecting logical fallacies has important applications in tracking misinformation and validating claims. In this paper, we design a process to reliably detect logical fallacies by translating natural language to First-order Logic (FOL) step-by-step using Large Language Models (LLMs). We then utilize Satisfiability Modulo Theory (SMT) solvers to reason about the validity of the formula and classify inputs as either a fallacy or valid statement. Our model also provides a novel means of utilizing LLMs to interpret the output of the SMT solver, offering insights into the counter-examples that illustrate why a given sentence is considered a logical fallacy. Our approach is robust, interpretable and does not require training data or fine-tuning. We evaluate our model on a mixed dataset of fallacies and valid sentences. The results demonstrate improved performance compared to end-to-end LLMs, with our classifier achieving an F1-score of 71\% on the Logic dataset. The approach is able to generalize effectively, achieving an F1-score of 73% on the challenge set, LogicClimate, outperforming state-of-the-art models by 21% despite its much smaller size.

In the arbitrary pattern formation problem, $n$ autonomous, mobile robots must form an arbitrary pattern $P \subseteq \mathbb{R}^2$. The (deterministic) robots are typically assumed to be indistinguishable, disoriented, and unable to communicate. An important distinction is whether robots have memory and/or a limited viewing range. Previous work managed to form $P$ under a natural symmetry condition if robots have no memory but an unlimited viewing range [22] or if robots have a limited viewing range but memory [25]. In the latter case, $P$ is only formed in a shrunk version that has constant diameter. Without memory and with limited viewing range, forming arbitrary patterns remains an open problem. We provide a partial solution by showing that $P$ can be formed under the same symmetry condition if the robots' initial diameter is $\leq 1$. Our protocol partitions $P$ into rotation-symmetric components and exploits the initial mutual visibility to form one cluster per component. Using a careful placement of the clusters and their robots, we show that a cluster can move in a coordinated way through its component while drawing $P$ by dropping one robot per pattern coordinate., Comment: 24 pages, 3 figures, submitted for SAND 2024, version with extended appendix

Instruction tuning is an important step in making language models useful for direct user interaction. However, many legal tasks remain out of reach for most open LLMs and there do not yet exist any large scale instruction datasets for the domain. This critically limits research in this application area. In this work, we curate LawInstruct, a large legal instruction dataset, covering 17 jurisdictions, 24 languages and a total of 12M examples. We present evidence that domain-specific pretraining and instruction tuning improve performance on LegalBench, including improving Flan-T5 XL by 8 points or 16\% over the baseline. However, the effect does not generalize across all tasks, training regimes, model sizes, and other factors. LawInstruct is a resource for accelerating the development of models with stronger information processing and decision making capabilities in the legal domain.

We introduce NeuroSynt, a neuro-symbolic portfolio solver framework for reactive synthesis. At the core of the solver lies a seamless integration of neural and symbolic approaches to solving the reactive synthesis problem. To ensure soundness, the neural engine is coupled with model checkers verifying the predictions of the underlying neural models. The open-source implementation of NeuroSynt provides an integration framework for reactive synthesis in which new neural and state-of-the-art symbolic approaches can be seamlessly integrated. Extensive experiments demonstrate its efficacy in handling challenging specifications, enhancing the state-of-the-art reactive synthesis solvers, with NeuroSynt contributing novel solves in the current SYNTCOMP benchmarks.

Numerous cutting-edge scientific technologies originate at the laboratory scale, but transitioning them to practical industry applications is a formidable challenge. Traditional pilot projects at intermediate scales are costly and time-consuming. An alternative, the E-pilot, relies on high-fidelity numerical simulations, but even these simulations can be computationally prohibitive at larger scales. To overcome these limitations, we propose a scalable, physics-constrained reduced order model (ROM) method. ROM identifies critical physics modes from small-scale unit components, projecting governing equations onto these modes to create a reduced model that retains essential physics details. We also employ Discontinuous Galerkin Domain Decomposition (DG-DD) to apply ROM to unit components and interfaces, enabling the construction of large-scale global systems without data at such large scales. This method is demonstrated on the Poisson and Stokes flow equations, showing that it can solve equations about $15 - 40$ times faster with only $\sim$ $1\%$ relative error. Furthermore, ROM takes one order of magnitude less memory than the full order model, enabling larger scale predictions at a given memory limitation., Comment: 40 pages, 12 figures. Submitted to Computer Methods in Applied Mechanics and Engineering

We present Self-Driven Strategy Learning ($\textit{sdsl}$), a lightweight online learning methodology for automated reasoning tasks that involve solving a set of related problems. $\textit{sdsl}$ does not require offline training, but instead automatically constructs a dataset while solving earlier problems. It fits a machine learning model to this data which is then used to adjust the solving strategy for later problems. We formally define the approach as a set of abstract transition rules. We describe a concrete instance of the sdsl calculus which uses conditional sampling for generating data and random forests as the underlying machine learning model. We implement the approach on top of the Kissat solver and show that the combination of Kissat+$\textit{sdsl}$ certifies larger bounds and finds more counter-examples than other state-of-the-art bounded model checking approaches on benchmarks obtained from the latest Hardware Model Checking Competition., Comment: Condensed version published at FMCAD'23

A rigorous formalization of desired system requirements is indispensable when performing any verification task. This often limits the application of verification techniques, as writing formal specifications is an error-prone and time-consuming manual task. To facilitate this, we present nl2spec, a framework for applying Large Language Models (LLMs) to derive formal specifications (in temporal logics) from unstructured natural language. In particular, we introduce a new methodology to detect and resolve the inherent ambiguity of system requirements in natural language: we utilize LLMs to map subformulas of the formalization back to the corresponding natural language fragments of the input. Users iteratively add, delete, and edit these sub-translations to amend erroneous formalizations, which is easier than manually redrafting the entire formalization. The framework is agnostic to specific application domains and can be extended to similar specification languages and new neural models. We perform a user study to obtain a challenging dataset, which we use to run experiments on the quality of translations. We provide an open-source implementation, including a web-based frontend.

We present a deep learning approach for repairing sequential circuits against formal specifications given in linear-time temporal logic (LTL). Given a defective circuit and its formal specification, we train Transformer models to output circuits that satisfy the corresponding specification. We propose a separated hierarchical Transformer for multimodal representation learning of the formal specification and the circuit. We introduce a data generation algorithm that enables generalization to more complex specifications and out-of-distribution datasets. In addition, our proposed repair mechanism significantly improves the automated synthesis of circuits from LTL specifications with Transformers. It improves the state-of-the-art by $6.8$ percentage points on held-out instances and $11.8$ percentage points on an out-of-distribution dataset from the annual reactive synthesis competition., Comment: To appear at ICLR'23

We analyze the convergence of the $k$-opinion Undecided State Dynamics (USD) in the population protocol model. For $k$=2 opinions it is well known that the USD reaches consensus with high probability within $O(n \log n)$ interactions. Proving that the process also quickly solves the consensus problem for $k>2$ opinions has remained open, despite analogous results for larger $k$ in the related parallel gossip model. In this paper we prove such convergence: under mild assumptions on $k$ and on the initial number of undecided agents we prove that the USD achieves plurality consensus within $O(k n \log n)$ interactions with high probability, regardless of the initial bias. Moreover, if there is an initial additive bias of at least $\Omega(\sqrt{n} \log n)$ we prove that the initial plurality opinion wins with high probability, and if there is a multiplicative bias the convergence time is further improved. Note that this is the first result for $k > 2$ for the USD in the population protocol model. Furthermore, it is the first result for the unsynchronized variant of the USD with $k>2$ which does not need any initial bias.

Hyperproperties relate multiple computation traces to each other. Model checkers for hyperproperties thus return, in case a system model violates the specification, a set of traces as a counterexample. Fixing the erroneous relations between traces in the system that led to the counterexample is a difficult manual effort that highly benefits from additional explanations. In this paper, we present an explanation method for counterexamples to hyperproperties described in the specification logic HyperLTL. We extend Halpern and Pearl's definition of actual causality to sets of traces witnessing the violation of a HyperLTL formula, which allows us to identify the events that caused the violation. We report on the implementation of our method and show that it significantly improves on previous approaches for analyzing counterexamples returned by HyperLTL model checkers., Comment: 34th International Conference on Computer-Aided Verification (CAV 2022)

We study the generalization abilities of language models when translating natural language into formal specifications with complex semantics. In particular, we fine-tune language models on three datasets consisting of English sentences and their corresponding formal representation: 1) regular expressions (regex), frequently used in programming and search; 2) First-order logic (FOL), commonly used in software verification and theorem proving; and 3) linear-time temporal logic (LTL), which forms the basis for industrial hardware specification languages. Our experiments show that, in these diverse domains, the language models maintain their generalization capabilities from pre-trained knowledge of natural language to generalize, e.g., to new variable names or operator descriptions. Additionally, they achieve competitive performance, and even outperform the state-of-the-art for translating into regular expressions, with the benefits of being easy to access, efficient to fine-tune, and without a particular need for domain-specific reasoning.

In this paper, we study the computation of how much an input token in a Transformer model influences its prediction. We formalize a method to construct a flow network out of the attention values of encoder-only Transformer models and extend it to general Transformer architectures including an auto-regressive decoder. We show that running a maxflow algorithm on the flow network construction yields Shapley values, which determine the impact of a player in cooperative game theory. By interpreting the input tokens in the flow network as players, we can compute their influence on the total attention flow leading to the decoder's decision. Additionally, we provide a library that computes and visualizes the attention flow of arbitrary Transformer models. We show the usefulness of our implementation on various models trained on natural language processing and reasoning tasks.

An enforcement mechanism monitors a reactive system for undesired behavior at runtime and corrects the system's output in case it violates the given specification. In this paper, we study the enforcement problem for hyperproperties, i.e., properties that relate multiple computation traces to each other. We elaborate the notion of sound and transparent enforcement mechanisms for hyperproperties in two trace input models: 1) the parallel trace input model, where the number of traces is known a-priori and all traces are produced and processed in parallel and 2) the sequential trace input model, where traces are processed sequentially and no a-priori bound on the number of traces is known. For both models, we study enforcement algorithms for specifications given as formulas in universally quantified HyperLTL, a temporal logic for hyperproperties. For the parallel model, we describe an enforcement mechanism based on parity games. For the sequential model, we show that enforcement is in general undecidable and present algorithms for reasonable simplifications of the problem (partial guarantees or the restriction to safety properties). Furthermore, we report on experimental results of our prototype implementation for the parallel model.

We study the capabilities of GANs and Wasserstein GANs equipped with Transformer encoders to generate sensible and challenging training data for symbolic reasoning domains. We conduct experiments on two problem domains where Transformers have been successfully applied recently: symbolic mathematics and temporal specifications in verification. Even without autoregression, our GAN models produce syntactically correct instances. We show that the generated data can be used as a substitute for real training data when training a classifier, and, especially, that training data can be generated from a dataset that is too small to be trained on directly. Using a GAN setting also allows us to alter the target distribution: We show that by adding a classifier uncertainty part to the generator objective, we obtain a dataset that is even harder to solve for a temporal logic classifier than our original dataset.

Model checkers provide algorithms for proving that a mathematical model of a system satisfies a given specification. In case of a violation, a counterexample that shows the erroneous behavior is returned. Understanding these counterexamples is challenging, especially for hyperproperty specifications, i.e., specifications that relate multiple executions of a system to each other. We aim to facilitate the visual analysis of such counterexamples through our HyperVis tool, which provides interactive visualizations of the given model, specification, and counterexample. Within an iterative and interdisciplinary design process, we developed visualization solutions that can effectively communicate the core aspects of the model checking result. Specifically, we introduce graphical representations of binary values for improving pattern recognition, color encoding for better indicating related aspects, visually enhanced textual descriptions, as well as extensive cross-view highlighting mechanisms. Further, through an underlying causal analysis of the counterexample, we are also able to identify values that contributed to the violation and use this knowledge for both improved encoding and highlighting. Finally, the analyst can modify both the specification of the hyperproperty and the system directly within HyperVis and initiate the model checking of the new version. In combination, these features notably support the analyst in understanding the error leading to the counterexample as well as iterating the provided system and specification. We ran multiple case studies with HyperVis and tested it with domain experts in qualitative feedback sessions. The participants' positive feedback confirms the considerable improvement over the manual, text-based status quo and the value of the tool for explaining hyperproperties., Comment: This work has been accepted for the IEEE VIS 2021 conference and will be published in the IEEE Transactions on Visualization and Computer Graphics journal

We train hierarchical Transformers on the task of synthesizing hardware circuits directly out of high-level logical specifications in linear-time temporal logic (LTL). The LTL synthesis problem is a well-known algorithmic challenge with a long history and an annual competition is organized to track the improvement of algorithms and tooling over time. New approaches using machine learning might open a lot of possibilities in this area, but suffer from the lack of sufficient amounts of training data. In this paper, we consider a method to generate large amounts of additional training data, i.e., pairs of specifications and circuits implementing them. We ensure that this synthetic data is sufficiently close to human-written specifications by mining common patterns from the specifications used in the synthesis competitions. We show that hierarchical Transformers trained on this synthetic data solve a significant portion of problems from the synthesis competitions, and even out-of-distribution examples from a recent case study.

We present a loosely-stabilizing phase clock for population protocols. In the population model we are given a system of $n$ identical agents which interact in a sequence of randomly chosen pairs. Our phase clock is leaderless and it requires $O(\log n)$ states. It runs forever and is, at any point of time, in a synchronous state w.h.p. When started in an arbitrary configuration, it recovers rapidly and enters a synchronous configuration within $O(n\log n)$ interactions w.h.p. Once the clock is synchronized, it stays in a synchronous configuration for at least poly $n$ parallel time w.h.p. We use our clock to design a loosely-stabilizing protocol that solves the comparison problem introduced by Alistarh et al., 2021. In this problem, a subset of agents has at any time either $A$ or $B$ as input. The goal is to keep track which of the two opinions is (momentarily) the majority. We show that if the majority has a support of at least $\Omega(\log n)$ agents and a sufficiently large bias is present, then the protocol converges to a correct output within $O(n\log n)$ interactions and stays in a correct configuration for poly $n$ interactions, w.h.p.

We studied the hyperlogic HyperQPTL, which combines the concepts of trace relations and $\omega$-regularity. We showed that HyperQPTL is very expressive, it can express properties like promptness, bounded waiting for a grant, epistemic properties, and, in particular, any $\omega$-regular property. Those properties are not expressible in previously studied hyperlogics like HyperLTL. At the same time, we argued that the expressiveness of HyperQPTL is optimal in a sense that a more expressive logic for $\omega$-regular hyperproperties would have an undecidable model checking problem. We furthermore studied the realizability problem of HyperQPTL. We showed that realizability is decidable for HyperQPTL fragments that contain properties like promptness. But still, in contrast to the satisfiability problem, propositional quantification does make the realizability problem of hyperlogics harder. More specifically, the HyperQPTL fragment of formulas with a universal-existential propositional quantifier alternation followed by a single trace quantifier is undecidable in general, even though the projection of the fragment to HyperLTL has a decidable realizability problem. Lastly, we implemented the bounded synthesis problem for HyperQPTL in the prototype tool BoSy. Using BoSy with HyperQPTL specifications, we have been able to synthesize several resource arbiters. The synthesis problem of non-linear-time hyperlogics is still open. For example, it is not yet known how to synthesize systems from specifications given in branching-time hyperlogics like HyperCTL$^*$., Comment: International Conference on Computer Aided Verification (CAV 2020)

Hyperproperties, such as non-interference and observational determinism, relate multiple computation traces with each other and are thus not monitorable by tools that consider computations in isolation. We present the monitoring approach implemented in the latest version of RVHyper, a runtime verification tool for hyperproperties. The input to the tool are specifications given in the temporal logic HyperLTL, which extends linear-time temporal logic (LTL) with trace quantifiers and trace variables. RVHyper processes execution traces sequentially until a violation of the specification is detected. In this case, a counter example, in the form of a set of traces, is returned. RVHyper employs a range of optimizations: a preprocessing analysis of the specification and a procedure that minimizes the traces that need to be stored during the monitoring process. In this article, we introduce a novel trace storage technique that arranges the traces in a tree-like structure to exploit partially equal traces. We evaluate RVhyper on existing benchmarks on secure information-flow control, error correcting codes and symmetry in hardware designs. As an example application outside of security, we show how RVHyper can be used to detect spurious dependencies in hardware designs., Comment: arXiv admin note: text overlap with arXiv:1807.00758, arXiv:1906.00798

Hyperproperties, which generalize trace properties by relating multiple traces, are widely studied in information-flow security. Recently, a number of logics for hyperproperties have been proposed, and there is a need to understand their decidability and relative expressiveness. The new logics have been obtained from standard logics with two principal extensions: temporal logics, like LTL and CTL$^*$, have been generalized to hyperproperties by adding variables for traces or paths. First-order and second-order logics, like monadic first-order logic of order and MSO, have been extended with the equal-level predicate. We study the impact of the two extensions across the spectrum of linear-time and branching-time logics, in particular for logics with quantification over propositions. The resulting hierarchy of hyperlogics differs significantly from the classical hierarchy, suggesting that the equal-level predicate adds more expressiveness than trace and path variables. Within the hierarchy of hyperlogics, we identify new boundaries on the decidability of the satisfiability problem. Specifically, we show that while HyperQPTL and HyperCTL$^*$ are both undecidable in general, formulas within their $\exists^*\forall^*$ fragments are decidable., Comment: Originally published at LICS 2019

We study two fundamental questions in neuro-symbolic computing: can deep learning tackle challenging problems in logics end-to-end, and can neural networks learn the semantics of logics. In this work we focus on linear-time temporal logic (LTL), as it is widely used in verification. We train a Transformer on the problem to directly predict a solution, i.e. a trace, to a given LTL formula. The training data is generated with classical solvers, which, however, only provide one of many possible solutions to each formula. We demonstrate that it is sufficient to train on those particular solutions to formulas, and that Transformers can predict solutions even to formulas from benchmarks from the literature on which the classical solver timed out. Transformers also generalize to the semantics of the logics: while they often deviate from the solutions found by the classical solvers, they still predict correct solutions to most formulas.