The development of vision-language models (VLMs) is driven by large-scale and diverse multimodal datasets. However, progress toward generalist biomedical VLMs is limited by the lack of annotated, publicly accessible datasets across biology and medicine. Existing efforts are restricted to narrow domains, missing the full diversity of biomedical knowledge encoded in scientific literature. To address this gap, we introduce BIOMEDICA, a scalable, open-source framework to extract, annotate, and serialize the entirety of the PubMed Central Open Access subset into an easy-to-use, publicly accessible dataset. Our framework produces a comprehensive archive with over 24 million unique image-text pairs from over 6 million articles. Metadata and expert-guided annotations are also provided. We demonstrate the utility and accessibility of our resource by releasing BMCA-CLIP, a suite of CLIP-style models continuously pre-trained on the BIOMEDICA dataset via streaming, eliminating the need to download 27 TB of data locally. On average, our models achieve state-of-the-art performance across 40 tasks - spanning pathology, radiology, ophthalmology, dermatology, surgery, molecular biology, parasitology, and cell biology - excelling in zero-shot classification with a 6.56% average improvement (as high as 29.8% and 17.5% in dermatology and ophthalmology, respectively), and stronger image-text retrieval, all while using 10x less compute. To foster reproducibility and collaboration, we release our codebase and dataset for the broader research community.

Multi-objective decision-making problems have emerged in numerous real-world scenarios, such as video games, navigation and robotics. Considering the clear advantages of Reinforcement Learning (RL) in optimizing decision-making processes, researchers have delved into the development of Multi-Objective RL (MORL) methods for solving multi-objective decision problems. However, previous methods either cannot obtain the entire Pareto front, or employ only a single policy network for all the preferences over multiple objectives, which may not produce personalized solutions for each preference. To address these limitations, we propose a novel decomposition-based framework for MORL, Pareto Set Learning for MORL (PSL-MORL), that harnesses the generation capability of hypernetwork to produce the parameters of the policy network for each decomposition weight, generating relatively distinct policies for various scalarized subproblems with high efficiency. PSL-MORL is a general framework, which is compatible for any RL algorithm. The theoretical result guarantees the superiority of the model capacity of PSL-MORL and the optimality of the obtained policy network. Through extensive experiments on diverse benchmarks, we demonstrate the effectiveness of PSL-MORL in achieving dense coverage of the Pareto front, significantly outperforming state-of-the-art MORL methods in the hypervolume and sparsity indicators.

As retrieval-augmented generation prevails in large language models, embedding models are becoming increasingly crucial. Despite the growing number of general embedding models, prior work often overlooks the critical role of training data quality. In this work, we introduce KaLM-Embedding, a general multilingual embedding model that leverages a large quantity of cleaner, more diverse, and domain-specific training data. Our model has been trained with key techniques proven to enhance performance: (1) persona-based synthetic data to create diversified examples distilled from LLMs, (2) ranking consistency filtering to remove less informative samples, and (3) semi-homogeneous task batch sampling to improve training efficacy. Departing from traditional BERT-like architectures, we adopt Qwen2-0.5B as the pre-trained model, facilitating the adaptation of auto-regressive language models for general embedding tasks. Extensive evaluations of the MTEB benchmark across multiple languages show that our model outperforms others of comparable size, setting a new standard for multilingual embedding models with less than 1B parameters.

The current evaluation of mathematical skills in LLMs is limited, as existing benchmarks are either relatively small, primarily focus on elementary and highschool problems, or lack diversity in topics. Additionally, the inclusion of visual elements in tasks remains largely under-explored. To address these gaps, we introduce U-MATH, a novel benchmark of 1,100 unpublished open-ended university-level problems sourced from teaching materials. It is balanced across six core subjects, with 20% of multimodal problems. Given the open-ended nature of U-MATH problems, we employ an LLM to judge the correctness of generated solutions. To this end, we release µ-MATH, a dataset to evaluate the LLMs' capabilities in judging solutions. The evaluation of general domain, math-specific, and multimodal LLMs highlights the challenges presented by U-MATH. Our findings reveal that LLMs achieve a maximum accuracy of only 63% on text-based tasks, with even lower 45% on visual problems. The solution assessment proves challenging for LLMs, with the best LLM judge having an F1-score of 80% on µ-MATH. Mathematical reasoning is a fundamental domain for assessing the true capabilities of Large Language Models (LLMs) to reason (Ahn et al., 2024). While existing benchmarks like GSM8K (Cobbe et al., 2021) and MATH (Hendrycks et al., 2021) provide valuable insights, they primarily focus on schoollevel mathematics. This leaves a significant gap in understanding how LLMs perform on more advanced, university-level problems.

In this paper we present an approach to reduce hallucinations in Large Language Models (LLMs) by incorporating Knowledge Graphs (KGs) as an additional modality. Our method involves transforming input text into a set of KG embeddings and using an adapter to integrate these embeddings into the language model space, without relying on external retrieval processes. To facilitate this, we created WikiEntities, a dataset containing over 3 million Wikipedia texts annotated with entities from Wikidata and their corresponding embeddings from PyTorch-BigGraph. This dataset serves as a valuable resource for training Entity Linking models and adapting the described method to various LLMs using specialized adapters. Our method does not require fine-tuning of the language models themselves; instead, we only train the adapter. This ensures that the model's performance on other tasks is not affected. We trained an adapter for the Mistral 7B, LLaMA 2-7B (chat), and LLaMA 3-8B (instruct) models using this dataset and demonstrated that our approach improves performance on the HaluEval, True-False benchmarks and FEVER dataset. The results indicate that incorporating KGs as a new modality can effectively reduce hallucinations and improve the factual accuracy of language models, all without the need for external retrieval.

The research related to digital twins has been increasing in recent years. Besides the mirroring of the physical word into the digital, there is the need of providing services related to the data collected and transferred to the virtual world. One of these services is the forecasting of physical part future behavior, that could lead to applications, like preventing harmful events or designing improvements to get better performance. One strategy used to predict any system operation it is the use of time series models like ARIMA or LSTM, and improvements were implemented using these algorithms. Recently, deep learning techniques based on generative models such as Generative Adversarial Networks (GANs) have been proposed to create time series and the use of LSTM has gained more relevance in time series forecasting, but both have limitations that restrict the forecasting results. Another issue found in the literature is the challenge of handling multivariate environments/applications in time series generation. Therefore, new methods need to be studied in order to fill these gaps and, consequently, provide better resources for creating useful digital twins. In this proposal, it is going to be studied the integration of a BiLSTM layer with a time series obtained by GAN in order to improve the forecasting of all the features provided by the dataset in terms of accuracy and, consequently, improving behaviour prediction.

Glioblastoma, a highly aggressive brain tumor, poses major challenges due to its poor prognosis and high morbidity rates. Partial differential equation-based models offer promising potential to enhance therapeutic outcomes by simulating patient-specific tumor behavior for improved radiotherapy planning. However, model calibration remains a bottleneck due to the high computational demands of optimization methods like Monte Carlo sampling and evolutionary algorithms. To address this, we recently introduced an approach leveraging a neural forward solver with gradient-based optimization to significantly reduce calibration time. This approach requires a highly accurate and fully differentiable forward model. We investigate multiple architectures, including (i) an enhanced TumorSurrogate, (ii) a modified nnU-Net, and (iii) a 3D Vision Transformer (ViT). The optimized TumorSurrogate achieved the best overall results, excelling in both tumor outline matching and voxel-level prediction of tumor cell concentration. It halved the MSE relative to the baseline model and achieved the highest Dice score across all tumor cell concentration thresholds. Our study demonstrates significant enhancement in forward solver performance and outlines important future research directions.

Hosted by Eleanor Drage and Kerry McInerney, The Good Robot is a podcast which explores the many complex intersections between gender, feminism and technology. In this episode, we talk to Sebastián Lehuedé, a Lecturer in Ethics, AI, and Society at King's College London. We talk about data activism in Chile, how water-intensive lithium extraction affects people living in the Atacama desert, the importance of reflexive research ethics, and an accidental Sunday afternoon shot of tequila. Sebastián's research focuses on the governance of digital technologies from a global social justice perspective. His current project, AI's Nature, explores the connection between Artificial Intelligence and environmental justice.

The global demand for unconventional energy sources such as geothermal energy and white hydrogen requires new exploration techniques for precise subsurface structure characterization and potential reservoir identification. The Magnetotelluric (MT) method is crucial for these tasks, providing critical information on the distribution of subsurface electrical resistivity at depths ranging from hundreds to thousands of meters. However, traditional iterative algorithm-based inversion methods require the adjustment of multiple parameters, demanding time-consuming and exhaustive tuning processes to achieve proper cost function minimization. Recent advances have incorporated deep learning algorithms for MT inversion, primarily based on supervised learning, and large labeled datasets are needed for training. This work utilizes TensorFlow operations to create a differentiable forward MT operator, leveraging its automatic differentiation capability. Moreover, instead of solving for the subsurface model directly, as classical algorithms perform, this paper presents a new deep unsupervised inversion algorithm guided by physics to estimate 1D MT models. Instead of using datasets with the observed data and their respective model as labels during training, our method employs a differentiable modeling operator that physically guides the cost function minimization, making the proposed method solely dependent on observed data. Therefore, the optimization algorithm updates the network weights to minimize the data misfit. We test the proposed method with field and synthetic data at different acquisition frequencies, demonstrating that the resistivity models obtained are more accurate than those calculated using other techniques.

We describe two Monte Carlo algorithms for sampling from the integrated posterior distributions of a range of Bayesian mixture models. Both algorithms allow us to directly sample not only the assignment of observations to components but also the number of components, thereby fitting the model and performing model selection over the number of components in a single computation. The first algorithm is a traditional collapsed Gibbs sampler, albeit with an unusual move-set; the second builds on the first, adding rejection-free sampling from the prior over component assignments, to create an algorithm that has excellent mixing time in typical applications and outperforms current state-of-the-art methods, in some cases by a wide margin. We demonstrate our methods with a selection of applications to latent class analysis.