The online evaluation of machine learning models is typically conducted through A/B experiments. Sequential statistical tests are valuable tools for analysing these experiments, as they enable researchers to stop data collection early without increasing the risk of false discoveries. However, existing sequential tests either limit the number of interim analyses or suffer from low statistical power. In this paper, we introduce a novel sequential test designed for the continuous monitoring of A/B experiments. We validate our method using semi-synthetic simulations and demonstrate that it outperforms current state-of-the-art sequential testing approaches. Our method is derived using a new technique that inverts a bound on the probability of threshold crossing, based on a classical maximal inequality.

A large number of empirical studies on applying self-attention models in the domain of recommender systems are based on offline evaluation and metrics computed on standardized datasets. Moreover, many of them do not consider side information such as item and customer metadata although deep-learning recommenders live up to their full potential only when numerous features of heterogeneous type are included. Also, normally the model is used only for a single use case. Due to these shortcomings, even if relevant, previous works are not always representative of their actual effectiveness in real-world industry applications. In this talk, we contribute to bridging this gap by presenting live experimental results demonstrating improvements in user retention of up to 30%. Moreover, we share our learnings and challenges from building a re-usable and configurable recommender system for various applications from the fashion industry. In particular, we focus on fashion inspiration use-cases, such as outfit ranking, outfit recommendation and real-time personalized outfit generation.

Memetic algorithms are a type of genetic algorithms very valuable in optimization problems. They are based on the concept of “meme”, and use local search techniques, which allow them to avoid premature convergence to suboptimal solutions. Among these algorithms we can consider Lamarckian and Baldwinian models, depending on whether they modify (the former) or not (the latter) the agent’s genotype. In this paper we analyze the application of memetic algorithms to the NPerson Iterated Prisoner’s Dilemma (NIPD). NIPD is an interesting game that has proved to be very useful to explore the emergence of cooperation in multi-player scenarios. The main contributions of this paper are related to setting the ground to understand the implications of the memetic model and the related parameters. We investigate to which extent these decisions determine the level of cooperation obtained as well as the memory and the execution performance.

In this article we tackle the problem of maximizing cooperation among self-interested agents in a resource exchange environment. Our main concern is the design of mechanisms for maximizing cooperation among self-interested agents in a way that their profits increase by exchanging or trading with resources. Although dynamic coalition formation and partner switching (rewiring) have been shown to promote the emergence and maintenance of cooperation for self-interested agents, no prior work in the literature has investigated whether merging both mechanisms exhibits positive synergies that lead to increase cooperation even further. Therefore, we introduce and analyze a novel dynamic coalition formation mechanism, that uses partner switching, to help self-interested agents to increase their profits in a resource exchange environment. Our experiments show the effectiveness of our mechanism at increasing the agents’ profits, as well as the emergence of trading as the preferred behavior over different types of complex networks.

The WiSafeCar (Wireless Traffic Safety Network between Cars) project aims at increasing the performance and reliability of the wireless transport and to provide traffic safety improvements. Within the context of this project, we have designed a Dynamic Carpooling System that will optimize the transport utilization by the ride sharing among people who usually cover the same route. An initial prototype of the system has been developed by using NetLogo. The information obtained from this simulator will be used to study the functioning of the clearing services, the current business models and to propose new ones. The first results seem encouraging, and the users have many economical advantages thanks to the sharing of costs which allows the individuals to retrench expenses and to contribute to the use of green technologies.

Evolutionary Programming (EP) seems a promising methodology to automatically find programs to solve new computing challenges. The Evolutionary Programming techniques use classical genetic operators (selection, crossover and mutation) to automatically generate programs targeted to solve computing problems or specifications. Among the methodologies related with Evolutionary Programming we can find Genetic Programming, Analytic Programming and Grammatical Evolution. In this paper we present the Evolutionary Programming Multiagent Systems (EPMAS) framework based on Grammatical Evolution (GE) to evolutionary generate Multi-agent systems (MAS) ad-hoc. We also present two case studies in MAS scenarios for applying our EPMAS framework: the predator-prey problem and the Iterative Prisoner's Dilemma. © ECMS.