Jorge Sánchez

Laura Alonso Alemany, Luciana Benotti, Hernán Maina et al.

Automated decision-making systems, especially those based on natural language processing, are pervasive in our lives. They are not only behind the internet search engines we use daily, but also take more critical roles: selecting candidates for a job, determining suspects of a crime, diagnosing autism and more. Such automated systems make errors, which may be harmful in many ways, be it because of the severity of the consequences (as in health issues) or because of the sheer number of people they affect. When errors made by an automated system affect a population more than others, we call the system \textit{biased}. Most modern natural language technologies are based on artifacts obtained from enormous volumes of text using machine learning, namely language models and word embeddings. Since they are created by applying subsymbolic machine learning, mostly artificial neural networks, they are opaque and practically uninterpretable by direct inspection, thus making it very difficult to audit them. In this paper, we present a methodology that spells out how social scientists, domain experts, and machine learning experts can collaboratively explore biases and harmful stereotypes in word embeddings and large language models. Our methodology is based on the following principles: * focus on the linguistic manifestations of discrimination on word embeddings and language models, not on the mathematical properties of the models * reduce the technical barrier for discrimination experts%, be it social scientists, domain experts or other * characterize through a qualitative exploratory process in addition to a metric-based approach * address mitigation as part of the training process, not as an afterthought

Jorge Sánchez, Guadalupe García-Isla, Sandra Perez-Herrero et al.

Designing optimizers that remain effective under tight evaluation budgets is critical in expensive black-box settings such as cardiac digital twinning. We propose Frenetic Cat-inspired Particle Optimization (FCPO), a hybrid swarm method that couples particle swarm optimization-like dynamics with an explicit-state Markov switching controller to schedule exploration and refinement operators online. FCPO integrates (i) state-conditioned bounded motion, (ii) an elite-difference global jump operator to escape stagnation, (iii) eigen-space guided local refinement from elite covariance, and (iv) linear population size reduction to control late-stage computational cost. We benchmark FCPO on five representative functions from the Congress on Evolutionary Computation (CEC) 2022 suite (F1, F2, F3, F6 and F10) at dimensions D$\in${10,20} over 30 independent runs, comparing against PSO, CSO, CLPSO, SHADE, L-SHADE and CMA-ES. FCPO achieves the lowest mean runtime across the ten benchmark cases (average 0.183 s), about 2.3x faster than CMA-ES and 2.6x faster than L-SHADE in our Python implementation. On the multimodal composition function F10 at D=20, FCPO attains the best mean objective (9.625x 10^2 $\pm$ 1.275x 10^3) and remains faster than CMA-ES (0.602 s vs. 1.126 s mean runtime). On structured landscapes (F1--F3) and on the hybrid function (F6), CMA-ES remains the most accurate method, while FCPO substantially improves over classical swarms and maintains a favorable accuracy--runtime trade-off. Finally, in a ventricular activation digital twin calibration task, FCPO reaches the target electrocardiogram (ECG) fidelity (RMSE<0.1 mV) within ~ 40 iterations and produces physiologically plausible activation maps with robust convergence across repeated initializations, supporting its use as a practical optimizer for expensive inverse problems.

Hernán Maina, Guido Ivetta, Mateo Lione Stuto et al.

Visually impaired people could benefit from Visual Question Answering (VQA) systems to interpret text in their surroundings. However, current models often struggle with recognizing text in the photos taken by this population. Through in-depth interviews with visually impaired individuals, we identified common framing conventions that frequently result in misaligned text. Existing VQA benchmarks primarily feature well-oriented text captured by sighted users, under-representing these challenges. To address this gap, we introduce ROtated SAmpling (ROSA), a decoding strategy that enhances VQA performance in text-rich images with incorrectly oriented text. ROSA outperforms Greedy decoding by 11.7 absolute points in the best-performing model.

Jorge Sánchez, Rodrigo Laguna

Cross-modal retrieval is the task of retrieving samples of a given modality by using queries of a different one. Due to the wide range of practical applications, the problem has been mainly focused on the vision and language case, e.g. text to image retrieval, where models like CLIP have proven effective in solving such tasks. The dominant approach to learning such coordinated representations consists of projecting them onto a common space where matching views stay close and those from non-matching pairs are pushed away from each other. Although this cross-modal coordination has been applied also to other pairwise combinations, extending it to an arbitrary number of diverse modalities is a problem that has not been fully explored in the literature. In this paper, we propose two different approaches to the problem. The first is based on an extension of the CLIP contrastive objective to an arbitrary number of input modalities, while the second departs from the contrastive formulation and tackles the coordination problem by regressing the cross-modal similarities towards a target that reflects two simple and intuitive constraints of the cross-modal retrieval task. We run experiments on two different datasets, over different combinations of input modalities and show that the approach is not only simple and effective but also allows for tackling the retrieval problem in novel ways. Besides capturing a more diverse set of pair-wise interactions, we show that we can use the learned representations to improve retrieval performance by combining the embeddings from two or more such modalities.

Felix Bach, Jochen Klar, Axel Loewe et al.

This work describes the setup of an advanced technical infrastructure for collaborative software development (CDE) in large, distributed projects based on GitLab. We present its customization and extension, additional features and processes like code review, continuous automated testing, DevOps practices, and sustainable life-cycle management including long-term preservation and citable publishing of software releases along with relevant metadata. The environment is currently used for developing the open cardiac simulation software openCARP and an evaluation showcases its capability and utility for collaboration and coordination of sizeable heterogeneous teams. As such, it could be a suitable and sustainable infrastructure solution for a wide range of research software projects.

Matías Molina, Jorge Sánchez

Zero-shot classification (ZSC) is the task of learning predictors for classes not seen during training. Although the different methods in the literature are evaluated using the same class splits, little is known about their stability under different class partitions. In this work we show experimentally that ZSC performance exhibits strong variability under changing training setups. We propose the use ensemble learning as an attempt to mitigate this phenomena.