Regularization-based approaches for injecting constraints in Machine Learning (ML) were introduced (see e.g. Given the recent interest in ethical and trustworthy AI, however, several works are resorting to these approaches for enforcing desired properties over a ML model (e.g. The regularization function C denotes a vector of (nonnegative) constraint violation indices for m constraints, while λ 0 is a vector of weights (or multipliers). As an example, in a regression problem we may desire a specific output ordering for two input vectors in the training set. If n is even, the term is 0 for perfectly balanced classifications.

Quantum computation is an emerging technology that promises a wide range of possible use cases. This promise is primarily based on algorithms that are unlikely to be viable over the coming decade. For near-term applications, quantum software needs to be carefully tailored to the hardware available. In this paper, we begin to explore whether near-term quantum computers could provide tools that are useful in the creation and implementation of computer games. The procedural generation of geopolitical maps and their associated history is considered as a motivating example. This is performed by encoding a rudimentary decision making process for the nations within a quantum procedure that is well-suited to near-term devices. Given the novelty of quantum computing within the field of procedural generation, we also provide an introduction to the basic concepts involved.

The Precedence Constrained Generalized Traveling Salesman Problem (PCGTSP) combines the Generalized Traveling Salesman Problem (GTSP) and the Sequential Ordering Problem (SOP). We present a novel branching technique for the GTSP which enables the extension of a powerful pruning technique. This is combined with some modifications of known bounding methods for related problems. The algorithm manages to solve problem instances with 12-26 groups within a minute, and instances with around 50 groups which are denser with precedence constraints within 24 hours.

This paper describes an extension to the DMN standard, called cDMN. It aims to enlarge the expressivity of DMN in order to solve more complex problems, while retaining DMN's goal of being readable by domain experts. We test cDMN by solving the most complex challenges posted on the DM Community website. We compare our own cDMN solutions to the solutions that have been submitted to the website and find that our approach is competitive, both in readability and compactness. Moreover, cDMN is able to solve more challenges than any other approach.

We can break symmetry by eliminating solutions within a symmetry class that are not least in the lexicographical ordering. This is often referred to as the lex-leader method. Unfortunately, as symmetry groups can be large, the lexleader method is not tractable in general. We prove that using other total orderings besides the usual lexicographical ordering will not reduce the computational complexity of breaking symmetry in general. It follows that breaking symmetry with other orderings like the Gray code ordering or the Snake-Lex ordering is intractable in general.

When learning to ride a bike, a child falls down a number of times before achieving the first success. As falling down usually has only mild consequences, it can be seen as a tolerable failure in exchange for a faster learning process, as it provides rich information about an undesired behavior. In the context of Bayesian optimization under unknown constraints (BOC), typical strategies for safe learning explore conservatively and avoid failures by all means. On the other side of the spectrum, non conservative BOC algorithms that allow failing may fail an unbounded number of times before reaching the optimum. In this work, we propose a novel decision maker grounded in control theory that controls the amount of risk we allow in the search as a function of a given budget of failures. Empirical validation shows that our algorithm uses the failures budget more efficiently in a variety of optimization experiments, and generally achieves lower regret, than state-of-the-art methods. In addition, we propose an original algorithm for unconstrained Bayesian optimization inspired by the notion of excursion sets in stochastic processes, upon which the failures-aware algorithm is built.

We address multi-armed bandits (MAB) where the objective is to maximize the cumulative reward under a probabilistic linear constraint. For a few real-world instances of this problem, constrained extensions of the well-known Thompson Sampling (TS) heuristic have recently been proposed. However, finite-time analysis of constrained TS is challenging; as a result, only O(\sqrt{T}) bounds on the cumulative reward loss (i.e., the regret) are available. In this paper, we describe LinConTS, a TS-based algorithm for bandits that place a linear constraint on the probability of earning a reward in every round. We show that for LinConTS, the regret as well as the cumulative constraint violations are upper bounded by O(\log T) for the suboptimal arms. We develop a proof technique that relies on careful analysis of the dual problem and combine it with recent theoretical work on unconstrained TS. Through numerical experiments on two real-world datasets, we demonstrate that LinConTS outperforms an asymptotically optimal upper confidence bound (UCB) scheme in terms of simultaneously minimizing the regret and the violation.

Linear integer constraints are one of the most important constraints in combinatorial problems since they are commonly found in many practical applications. Typically, encodings to Boolean satisfiability (SAT) format of conjunctive normal form perform poorly in problems with these constraints in comparison with SAT modulo theories (SMT), lazy clause generation (LCG) or mixed integer programming (MIP) solvers. In this paper we explore and categorize SAT encodings for linear integer constraints. We define new SAT encodings based on multi-valued decision diagrams, and sorting networks. We compare different SAT encodings of linear constraints and demonstrate where one may be preferable to another. We also compare SAT encodings against other solving methods and show they can be better than linear integer (MIP) solvers and sometimes better than LCG or SMT solvers on appropriate problems. Combining the new encoding with lazy decomposition, which during runtime only encodes constraints that are important to the solving process that occurs, gives the best option for many highly combinatorial problems involving linear constraints.

The Traveling Salesman Problem is a well known challenge in Computer Science: it consists on finding the shortest route possible that traverses all cities in a given map only once. Although its simple explanation, this problem is, indeed, NP-Complete. This implies that the difficulty to solve it increases rapidly with the number of cities, and we do not know in fact a general solution that solves the problem. For that reason, we currently consider that any method able to find a sub-optimal solution is generally good enough (we cannot verify if the solution returned is the optimal one most of the times). To solve it, we can try to apply a modification of the Self-Organizing Map (SOM) technique.

Contextual multi-armed bandit (MAB) achieves cutting-edge performance on a variety of problems. When it comes to real-world scenarios such as recommendation system and online advertising, however, it is essential to consider the resource consumption of exploration. In practice, there is typically non-zero cost associated with executing a recommendation (arm) in the environment, and hence, the policy should be learned with a fixed exploration cost constraint. It is challenging to learn a global optimal policy directly, since it is a NP-hard problem and significantly complicates the exploration and exploitation trade-off of bandit algorithms. Existing approaches focus on solving the problems by adopting the greedy policy which estimates the expected rewards and costs and uses a greedy selection based on each arm's expected reward/cost ratio using historical observation until the exploration resource is exhausted. However, existing methods are hard to extend to infinite time horizon, since the learning process will be terminated when there is no more resource. In this paper, we propose a hierarchical adaptive contextual bandit method (HATCH) to conduct the policy learning of contextual bandits with a budget constraint. HATCH adopts an adaptive method to allocate the exploration resource based on the remaining resource/time and the estimation of reward distribution among different user contexts. In addition, we utilize full of contextual feature information to find the best personalized recommendation. Finally, in order to prove the theoretical guarantee, we present a regret bound analysis and prove that HATCH achieves a regret bound as low as $O(\sqrt{T})$. The experimental results demonstrate the effectiveness and efficiency of the proposed method on both synthetic data sets and the real-world applications.