Improving active Mealy machine learning for protocol conformance testing

Using a well-known industrial case study from the verification literature, the bounded retransmission protocol, we show how active learning can be used to establish the correctness of protocol implementation I relative to a given reference implementation R. Using active learning, we learn a model M _R of reference implementation R, which serves as input for a model-based testing tool that checks conformance of implementation I to M _R . In addition, we also explore an alternative approach in which we learn a model M _I of implementation I, which is compared to model M _R using an equivalence checker. Our work uses a unique combination of software tools for model construction (Uppaal), active learning (LearnLib, Tomte), model-based testing (JTorX, TorXakis) and verification (CADP, MRMC). We show how these tools can be used for learning models of and revealing errors in implementations, present the new notion of a conformance oracle, and demonstrate how conformance oracles can be used to speed up conformance checking.     

Three naive Bayes approaches for discrimination-free classification

In this paper, we investigate how to modify the naive Bayes classifier in order to perform classification that is restricted to be independent with respect to a given sensitive attribute. Such independency restrictions occur naturally when the decision process leading to the labels in the data-set was biased; e.g., due to gender or racial discrimination. This setting is motivated by many cases in which there exist laws that disallow a decision that is partly based on discrimination. Naive application of machine learning techniques would result in huge fines for companies. We present three approaches for making the naive Bayes classifier discrimination-free: (i) modifying the probability of the decision being positive, (ii) training one model for every sensitive attribute value and balancing them, and (iii) adding a latent variable to the Bayesian model that represents the unbiased label and optimizing the model parameters for likelihood using expectation maximization. We present experiments for the three approaches on both artificial and real-life data.     

The efficiency of identifying timed automata and the power of clocks

We develop theory on the efficiency of identifying (learning) timed automata. In particular, we show that: (i) deterministic timed automata cannot be identified efficiently in the limit from labeled data and (ii) that one-clock deterministic timed automata can be identified efficiently in the limit from labeled data. We prove these results based on the distinguishability of these classes of timed automata. More specifically, we prove that the languages of deterministic timed automata cannot, and that one-clock deterministic timed automata can be distinguished from each other using strings in length bounded by a polynomial. In addition, we provide an algorithm that identifies one-clock deterministic timed automata efficiently from labeled data.Our results have interesting consequences for the power of clocks that are interesting also out of the scope of the identification problem.     

Black-box combinatorial optimization using models with integer-valued minima

Annals of Mathematics and Artificial Intelligence - When a black-box optimization objective can only be evaluated with costly or noisy measurements, most standard optimization algorithms are...     

Effect of the melting temperature on the crystallization behavior of a poly(l‐lactide)/poly(d‐lactide) equimolar mixture

The effect of the final melting temperature (T_f) on the crystallization of poly(l ‐lactide) (PLLA)/poly(d ‐lactide) (PDLA) was studied via a combination of differential scanning calorimetry, wide‐angle X‐ray scattering, polarized optical microscopy, and Fourier transform infrared (FTIR) spectroscopy. We observed that a residual stereocomplex (SC) crystal induced the formation of SC crystals during cooling from a T_f (230°C) just above the melting peak of the SC crystals. On cooling from a T_f (240°C) just above the endset temperature of SC crystal melting [T_m(S)(E)], the possible order structure and the strong interchain interaction promoted the preferential crystallization of SC crystals; this enhanced the formation of α crystals. During cooling from a T_f (≥250°C) far above T_m(S)(E), the crystallization peaks of α and SC crystals converged. The FTIR results indicated that the residual SC crystals, possible ordered structure, and interchain interactions in the melt might have been the key factors for the different crystallization of PLLA/PDLA. © 2015 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2016 , 133, 43015.     

Structural analogy between β′ triacylglycerols and n-alkanes. Toward the crystal structure of β′-2 p.p+2.p triacylglycerols

The structural elements of the β′-2 polymorph of 1,3-dilauroyl-2-myristoylglycerol as found by Birker et al. (J. Am. Oil Chem. Soc. 68:895–906, 1991) were also observed in the crystal structures of other long-chain compounds. This analogy led to the assembly of a β′-2 structure at the atomic level from known crystallographic data. The structure was optimized by molecular mechanics and was consistent with experimental data, including satisfactory reproduction of the X-ray powder pattern. To the best of our knowledge, this is the first β′ structure with a 1,2 configuration and an intramolecular orthorhombic subcell which is fully optimized by molecular mechanics to date. It shows all structural elements found earlier by Birker et al.     

A study of B-convergence of Runge-Kutta methods

Computing - This paper deals with the convergence analysis of implicit Runge-Kutta methods as applied to stiff, semilinear systems of the form $$\dot U$$ (t)=QU(t)+g(t, U(t)). A criterion is...     

Effects of cyclosporin A withdrawal on renal function and renal stimulation in liver transplant patients treated with triple-drug immunosuppression for over 2 years

The influence of CsA withdrawal on the glomerular filtration rate (GFR) and the effective renal plasma flow (ERPF) was prospectively studied in nine stable liver transplant recipients. Before CsA withdrawal (test 1), and 6 months thereafter (test 2) the renal function was determined by measuring GFR and the ERPF with 125I-iothalamate and 131I-hippuran respectively. The renal function was also stimulated with dopamine, with an amino-acid infusion and a combination of both. After CsA withdrawal the GFR increased, median from 74 ml min-1 to 90 ml min-1, (P < 0.04). The ERPF also increased, median from 310 ml min-1 to 380 ml min-1, (P < 0.03). In test 1 as well as in test 2 the renal function could be stimulated, especially with dopamine. GFR and ERPF improved, even after more than 2 years of CsA treatment. These results suggest that long-term CsA treatment impairs the renal function, though in these liver transplant patients CsA treatment did not prevent afferent and efferent arteriolar vasodilatation after renal stimulation. This reversible intrarenal vasoconstriction during CsA treatment may predict renal improvement after CsA withdrawal.     

One-step splitting methods for semi-discrete parabolic equations

The main purpose of the paper is to discuss splitting methods for parabolic equations via the method of lines. Firstly, we deal with the formulation of these methods for autonomous semi-discrete equations $$\frac{{dy}}{{dt}} = f(y),{\rm E}f{\rm E}non - linear,$$ f satisfying a linear splitting relation \(f(y) = \sum\limits_{i = 1}^k {f_i (y)} \) . A class of one-step integration formulas is defined, which is shown to contain all known splitting methods, provided the functionsf _i are defined appropriately. For a number of methods stability results are given. Secondly, attention is paid to alternating direction methods for problems with an arbitrary non-linear coupling between space derivatives.