We introduce VideoComp, a benchmark and learning framework for advancing video-text compositionality understanding, aimed at improving vision-language models (VLMs) in fine-grained temporal alignment. Unlike existing benchmarks focused on static image-text compositionality or isolated single-event videos, our benchmark targets alignment in continuous multi-event videos. Leveraging video-text datasets with temporally localized event captions (e.g. ActivityNet-Captions, YouCook2), we construct two compositional benchmarks, ActivityNet-Comp and YouCook2-Comp. We create challenging negative samples with subtle temporal disruptions such as reordering, action word replacement, partial captioning, and combined disruptions. These benchmarks comprehensively test models' compositional sensitivity across extended, cohesive video-text sequences. To improve model performance, we propose a hierarchical pairwise preference loss that strengthens alignment with temporally accurate pairs and gradually penalizes increasingly disrupted ones, encouraging fine-grained compositional learning. To mitigate the limited availability of densely annotated video data, we introduce a pretraining strategy that concatenates short video-caption pairs to simulate multi-event sequences. We evaluate video-text foundational models and large multimodal models (LMMs) on our benchmark, identifying both strengths and areas for improvement in compositionality. Overall, our work provides a comprehensive framework for evaluating and enhancing model capabilities in achieving fine-grained, temporally coherent video-text alignment.
Multiple studies have performed next-day fire prediction using satellite imagery. Two main satellites are used to detect wildfires: MODIS and VIIRS. Both satellites provide fire mask products, called MOD14 and VNP14, respectively. Studies have used one or the other, but there has been no comparison between them to determine which might be more suitable for next-day fire prediction. In this paper, we first evaluate how well VIIRS and MODIS data can be used to forecast wildfire spread one day ahead. We find that the model using VIIRS as input and VNP14 as target achieves the best results. Interestingly, the model using MODIS as input and VNP14 as target performs significantly better than using VNP14 as input and MOD14 as target. Next, we discuss why MOD14 might be harder to use for predicting next-day fires. We find that the MOD14 fire mask is highly stochastic and does not correlate with reasonable fire spread patterns. This is detrimental for machine learning tasks, as the model learns irrational patterns. Therefore, we conclude that MOD14 is unsuitable for next-day fire prediction and that VNP14 is a much better option. However, using MODIS input and VNP14 as target, we achieve a significant improvement in predictability. This indicates that an improved fire detection model is possible for MODIS. The full code and dataset is available online: https://github.com/justuskarlsson/wildfire-mod14-vnp14
Accurate surgical instrument segmentation is essential in cataract surgery for tasks such as skill assessment and workflow optimization. However, limited annotated data makes it difficult to develop fully automatic models. Prompt-based methods like SAM2 offer flexibility yet remain highly sensitive to the point prompt placement, often leading to inconsistent segmentations. We address this issue by introducing RP-SAM2, which incorporates a novel shift block and a compound loss function to stabilize point prompts. Our approach reduces annotator reliance on precise point positioning while maintaining robust segmentation capabilities. Experiments on the Cataract1k dataset demonstrate that RP-SAM2 improves segmentation accuracy, with a 2% mDSC gain, a 21.36% reduction in mHD95, and decreased variance across random single-point prompt results compared to SAM2. Additionally, on the CaDIS dataset, pseudo masks generated by RP-SAM2 for fine-tuning SAM2's mask decoder outperformed those generated by SAM2. These results highlight RP-SAM2 as a practical, stable and reliable solution for semi-automatic instrument segmentation in data-constrained medical settings. The code is available at https://github.com/BioMedIA-MBZUAI/RP-SAM2.
Financial scenario simulation is essential for risk management and portfolio optimization, yet it remains challenging especially in high-dimensional and small data settings common in finance. We propose a diffusion factor model that integrates latent factor structure into generative diffusion processes, bridging econometrics with modern generative AI to address the challenges of the curse of dimensionality and data scarcity in financial simulation. By exploiting the low-dimensional factor structure inherent in asset returns, we decompose the score function--a key component in diffusion models--using time-varying orthogonal projections, and this decomposition is incorporated into the design of neural network architectures. We derive rigorous statistical guarantees, establishing nonasymptotic error bounds for both score estimation at O(d^{5/2} n^{-2/(k+5)}) and generated distribution at O(d^{5/4} n^{-1/2(k+5)}), primarily driven by the intrinsic factor dimension k rather than the number of assets d, surpassing the dimension-dependent limits in the classical nonparametric statistics literature and making the framework viable for markets with thousands of assets. Numerical studies confirm superior performance in latent subspace recovery under small data regimes. Empirical analysis demonstrates the economic significance of our framework in constructing mean-variance optimal portfolios and factor portfolios. This work presents the first theoretical integration of factor structure with diffusion models, offering a principled approach for high-dimensional financial simulation with limited data. Our code is available at https://github.com/xymmmm00/diffusion_factor_model.
Cancer remains a leading global health challenge and a major cause of mortality. This study leverages machine learning (ML) to predict the survivability of cancer patients with metastatic patterns using the comprehensive MSK-MET dataset, which includes genomic and clinical data from 25,775 patients across 27 cancer types. We evaluated five ML models-XGBoost, Na\"ive Bayes, Decision Tree, Logistic Regression, and Random Fores using hyperparameter tuning and grid search. XGBoost emerged as the best performer with an area under the curve (AUC) of 0.82. To enhance model interpretability, SHapley Additive exPlanations (SHAP) were applied, revealing key predictors such as metastatic site count, tumor mutation burden, fraction of genome altered, and organ-specific metastases. Further survival analysis using Kaplan-Meier curves, Cox Proportional Hazards models, and XGBoost Survival Analysis identified significant predictors of patient outcomes, offering actionable insights for clinicians. These findings could aid in personalized prognosis and treatment planning, ultimately improving patient care.
We propose an enhanced deep hedging framework for index option portfolios, grounded in a realistic market simulator that captures the joint dynamics of S&P 500 returns and the full implied volatility surface. Our approach integrates surface-informed decisions with multiple hedging instruments and explicitly accounts for transaction costs. The hedging strategy also considers the variance risk premium embedded in the hedging instruments, enabling more informed and adaptive risk management. In this setting, state-dependent no-trade regions emerge naturally, improving rebalancing efficiency and hedging performance. Tested across simulated and historical data from 1996 to 2020, our method consistently outperforms traditional delta and delta-gamma hedging, demonstrating superior adaptability and risk reduction.
Accurate prediction of compound-protein interactions (CPI) remains a cornerstone challenge in computational drug discovery. While existing sequence-based approaches leverage molecular fingerprints or graph representations, they critically overlook three-dimensional (3D) structural determinants of binding affinity. To bridge this gap, we present EquiCPI, an end-to-end geometric deep learning framework that synergizes first-principles structural modeling with SE(3)-equivariant neural networks. Our pipeline transforms raw sequences into 3D atomic coordinates via ESMFold for proteins and DiffDock-L for ligands, followed by physics-guided conformer re-ranking and equivariant feature learning. At its core, EquiCPI employs SE(3)-equivariant message passing over atomic point clouds, preserving symmetry under rotations, translations, and reflections, while hierarchically encoding local interaction patterns through tensor products of spherical harmonics. The proposed model is evaluated on BindingDB (affinity prediction) and DUD-E (virtual screening), EquiCPI achieves performance on par with or exceeding the state-of-the-art deep learning competitors.
We present a general computational framework for solving continuous-time financial market equilibria under minimal modeling assumptions while incorporating realistic financial frictions, such as trading costs, and supporting multiple interacting agents. Inspired by generative adversarial networks (GANs), our approach employs a novel generative deep reinforcement learning framework with a decoupling feedback system embedded in the adversarial training loop, which we term as the \emph{reinforcement link}. This architecture stabilizes the training dynamics by incorporating feedback from the discriminator. Our theoretically guided feedback mechanism enables the decoupling of the equilibrium system, overcoming challenges that hinder conventional numerical algorithms. Experimentally, our algorithm not only learns but also provides testable predictions on how asset returns and volatilities emerge from the endogenous trading behavior of market participants, where traditional analytical methods fall short. The design of our model is further supported by an approximation guarantee.
Outlier detection in high-dimensional tabular data is challenging since data is often distributed across multiple lower-dimensional subspaces -- a phenomenon known as the Multiple Views effect (MV). This effect led to a large body of research focused on mining such subspaces, known as subspace selection. However, as the precise nature of the MV effect was not well understood, traditional methods had to rely on heuristic-driven search schemes that struggle to accurately capture the true structure of the data. Properly identifying these subspaces is critical for unsupervised tasks such as outlier detection or clustering, where misrepresenting the underlying data structure can hinder the performance. We introduce Myopic Subspace Theory (MST), a new theoretical framework that mathematically formulates the Multiple Views effect and writes subspace selection as a stochastic optimization problem. Based on MST, we introduce V-GAN, a generative method trained to solve such an optimization problem. This approach avoids any exhaustive search over the feature space while ensuring that the intrinsic data structure is preserved. Experiments on 42 real-world datasets show that using V-GAN subspaces to build ensemble methods leads to a significant increase in one-class classification performance -- compared to existing subspace selection, feature selection, and embedding methods. Further experiments on synthetic data show that V-GAN identifies subspaces more accurately while scaling better than other relevant subspace selection methods. These results confirm the theoretical guarantees of our approach and also highlight its practical viability in high-dimensional settings.
Phylogenetic trees are simple models of evolutionary processes. They describe conditionally independent divergent evolution of taxa from common ancestors. Phylogenetic trees commonly do not have enough flexibility to adequately model all evolutionary processes. For example, introgressive hybridization, where genes can flow from one taxon to another. Phylogenetic networks model evolution not fully described by a phylogenetic tree. However, many phylogenetic network models assume ancestral taxa merge instantaneously to form ``hybrid'' descendant taxa. In contrast, our convergence-divergence models retain a single underlying ``principal'' tree, but permit gene flow over arbitrary time frames. Alternatively, convergence-divergence models can describe other biological processes leading to taxa becoming more similar over a time frame, such as replicated evolution. Here we present novel maximum likelihood-based algorithms to infer most aspects of $N$-taxon convergence-divergence models, many consistently, using a quartet-based approach. The algorithms can be applied to multiple sequence alignments restricted to genes or genomic windows or to gene presence/absence datasets.
There are many areas of scientific endeavour where large, complex datasets are needed for benchmarking. Evolutionary computing provides a means towards creating such sets. As a case study, we consider Conway's Surreal numbers. They have largely been treated as a theoretical construct, with little effort towards empirical study, at least in part because of the difficulty of working with all but the smallest numbers. To advance this status, we need efficient algorithms, and in order to develop such we need benchmark data sets of surreal numbers. In this paper, we present a method for generating ensembles of random surreal numbers to benchmark algorithms. The approach uses an evolutionary algorithm to create the benchmark datasets where we can analyse and control features of the resulting test sets. Ultimately, the process is designed to generate networks with defined properties, and we expect this to be useful for other types of network data.
Continual learning in large language models (LLMs) is prone to catastrophic forgetting, where adapting to new tasks significantly degrades performance on previously learned ones. Existing methods typically rely on low-rank, parameter-efficient updates that limit the model's expressivity and introduce additional parameters per task, leading to scalability issues. To address these limitations, we propose a novel continual full fine-tuning approach leveraging adaptive singular value decomposition (SVD). Our method dynamically identifies task-specific low-rank parameter subspaces and constrains updates to be orthogonal to critical directions associated with prior tasks, thus effectively minimizing interference without additional parameter overhead or storing previous task gradients. We evaluate our approach extensively on standard continual learning benchmarks using both encoder-decoder (T5-Large) and decoder-only (LLaMA-2 7B) models, spanning diverse tasks including classification, generation, and reasoning. Empirically, our method achieves state-of-the-art results, up to 7% higher average accuracy than recent baselines like O-LoRA, and notably maintains the model's general linguistic capabilities, instruction-following accuracy, and safety throughout the continual learning process by reducing forgetting to near-negligible levels. Our adaptive SVD framework effectively balances model plasticity and knowledge retention, providing a practical, theoretically grounded, and computationally scalable solution for continual learning scenarios in large language models.
Hierarchical inductive biases are hypothesized to promote generalizable policies in reinforcement learning, as demonstrated by explicit hyperbolic latent representations and architectures. Therefore, a more flexible approach is to have these biases emerge naturally from the algorithm. We introduce Free Random Projection, an input mapping grounded in free probability theory that constructs random orthogonal matrices where hierarchical structure arises inherently. The free random projection integrates seamlessly into existing in-context reinforcement learning frameworks by encoding hierarchical organization within the input space without requiring explicit architectural modifications. Empirical results on multi-environment benchmarks show that free random projection consistently outperforms the standard random projection, leading to improvements in generalization. Furthermore, analyses within linearly solvable Markov decision processes and investigations of the spectrum of kernel random matrices reveal the theoretical underpinnings of free random projection's enhanced performance, highlighting its capacity for effective adaptation in hierarchically structured state spaces.
Predicting competitive outcomes typically requires fitting dynamical models to data, from which interaction strengths and coexistence indicators such as invasion criteria can be produced. Methods that allow to propagate parameter uncertainty are particularly indicated. These should ideally allow for competition between and within species at various life-stages, and make the best out of multiple data sources, each of which can be relatively scarce by statistical standards. Here, we embed a mathematical model of stage-structured competition between two species, producing analytical invasion criteria, into a two-species Integrated Population Model. The community-level IPM allows to combine counts, capture-recapture, and fecundity data into a single statistical framework, and the Bayesian formulation of the IPM fully propagates parameter uncertainty into invasion criteria. Model fitting demonstrates that we can correctly predict coexistence through reciprocal invasion when present, but that interaction strengths are not always estimable, depending on the prior chosen. Our competitive exclusion scenario is shown to be harder to identify, although our model allows to at least flag this scenario as uncertain rather than mistakenly present it as coexistence. Our results confirm the importance of accounting for uncertainty in the prediction of competitive outcomes.
In genome-scale constraint-based metabolic models, gene deletion strategies are crucial for achieving growth-coupled production, where cell growth and target metabolite production are simultaneously achieved. While computational methods for calculating gene deletions have been widely explored and contribute to developing gene deletion strategy databases, current approaches are limited in leveraging new data-driven paradigms, such as machine learning, for more efficient strain design. Therefore, it is necessary to propose a fundamental framework for this objective. In this study, we first formulate the problem of gene deletion strategy prediction and then propose a framework for predicting gene deletion strategies for growth-coupled production in genome-scale metabolic models. The proposed framework leverages deep learning algorithms to learn and integrate sequential gene and metabolite data representation, enabling the automatic gene deletion strategy prediction. Computational experiment results demonstrate the feasibility of the proposed framework, showing substantial improvements over baseline methods. Specifically, the proposed framework achieves a 14.69%, 22.52%, and 13.03% increase in overall accuracy across three metabolic models of different scales under study, while maintaining balanced precision and recall in predicting gene deletion statuses. The source code and examples for the framework are publicly available at https://github.com/MetNetComp/DeepGDel.
Motivation: Viruses represent the most abundant biological entities on the planet and play vital roles in diverse ecosystems. Cataloging viruses across various environments is essential for understanding their properties and functions. Metagenomic sequencing has emerged as the most comprehensive method for virus discovery, enabling the sequencing of all genetic materials, including viruses, from host or environmental samples. However, distinguishing viral sequences from the vast background of cellular organism-derived reads in metagenomic data remains a significant challenge. While several learning-based tools, such as VirSorter2 and geNomad, have shown promise in identifying viral contigs, they often experience varying degrees of false positive rates due to noise in sequencing and assembly, shared genes between viruses and their hosts, and the formation of proviruses within host genomes. This highlights the urgent need for an accurate and efficient method to evaluate the quality of viral contigs. Results: To address these challenges, we introduce ViralQC, a tool designed to assess the quality of reported viral contigs or bins. ViralQC identifies contamination regions within putative viral sequences using foundation models trained on viral and cellular genomes and estimates viral completeness through protein organization alignment. We evaluate ViralQC on multiple datasets and compare its performance against CheckV, the state-of-the-art in virus quality assessment. Notably, ViralQC correctly identifies 38% more contamination than CheckV, while maintaining a median absolute error of only 3%. In addition, ViralQC delivers more accurate results for medium- to high-quality (>50% completeness) contigs, demonstrating its superior performance in completeness estimation.
Recently, linear regression models incorporating an optimal transport (OT) loss have been explored for applications such as supervised unmixing of spectra, music transcription, and mass spectrometry. However, these task-specific approaches often do not generalize readily to a broader class of linear models. In this work, we propose a novel algorithmic framework for solving a general class of non-negative linear regression models with an entropy-regularized OT datafit term, based on Sinkhorn-like scaling iterations. Our framework accommodates convex penalty functions on the weights (e.g. squared-$\ell_2$ and $\ell_1$ norms), and admits additional convex loss terms between the transported marginal and target distribution (e.g. squared error or total variation). We derive simple multiplicative updates for common penalty and datafit terms. This method is suitable for large-scale problems due to its simplicity of implementation and straightforward parallelization.
The worldwide economy and environmental sustainability depend on eff icient and reliable supply chains, in which container shipping plays a crucial role as an environmentally friendly mode of transport. Liner shipping companies seek to improve operational efficiency by solving the stowage planning problem. Due to many complex combinatorial aspects, stowage planning is challenging and often decomposed into two NP-hard subproblems: master and slot planning. This article proposes AI2STOW, an end-to-end deep reinforcement learning model with feasibility projection and an action mask to create master plans under demand uncertainty with global objectives and constraints, including paired block stowage patterms. Our experimental results demonstrate that AI2STOW outperforms baseline methods from reinforcement learning and stochastic programming in objective performance and computational efficiency, based on simulated instances reflecting the scale of realistic vessels and operational planning horizons.
Covert speech involves imagining speaking without audible sound or any movements. Decoding covert speech from electroencephalogram (EEG) is challenging due to a limited understanding of neural pronunciation mapping and the low signal-to-noise ratio of the signal. In this study, we developed a large-scale multi-utterance speech EEG dataset from 57 right-handed native English-speaking subjects, each performing covert and overt speech tasks by repeating the same word in five utterances within a ten-second duration. Given the spatio-temporal nature of the neural activation process during speech pronunciation, we developed a Functional Areas Spatio-temporal Transformer (FAST), an effective framework for converting EEG signals into tokens and utilizing transformer architecture for sequence encoding. Our results reveal distinct and interpretable speech neural features by the visualization of FAST-generated activation maps across frontal and temporal brain regions with each word being covertly spoken, providing new insights into the discriminative features of the neural representation of covert speech. This is the first report of such a study, which provides interpretable evidence for speech decoding from EEG. The code for this work has been made public at https://github.com/Jiang-Muyun/FAST
A new benchmark dataset for functional data analysis (FDA) is presented, focusing on the reconstruction of eye movements from EEG data. The contribution is twofold: first, open challenges and evaluation metrics tailored to FDA applications are proposed. Second, functional neural networks are used to establish baseline results for the primary regression task of reconstructing eye movements from EEG signals. Baseline results are reported for the new dataset, based on consumer-grade hardware, and the EEGEyeNet dataset, based on research-grade hardware.