In survival analysis, traditional models assume all individuals will eventually experience the event of interest. However, advances in therapeutics have led to multiple clinical contexts with potentially curative therapies, and in these contexts, certain individuals may never experience the event. Statisticians have developed cure models as a methodology to address this challenge. Nonetheless, despite significant statistical advances in cure models, we have seen more limited uptake in biomedical applications, and we hypothesize that this is caused by limited guidance in the appropriate application of cure models. Cure models require specific identifiability conditions for valid parameter estimation, and previous reports have demonstrated significant issues with the inappropriate application of cure models. Existing tutorials for cure models focus on model implementation and either assume or provide only limited guidance on whether cure modeling is appropriate for the given dataset. This tutorial addresses this gap by describing a systematic procedure that integrates clinical judgment, visual inspection of Kaplan-Meier curves, and quantitative evaluation. We provide a worked example
Understanding propagation mechanisms in complex networks is essential for fields like epidemiology and multi-robot networks. This paper reviews various propagation models, from traditional deterministic frameworks to advanced data-driven and deep learning approaches. We differentiate between static and dynamic networks, noting that static models provide foundational insights, while dynamic models capture real-world temporal changes. Deterministic models like the SIR framework offer clear mathematical insights but often lack adaptability to randomness, whereas stochastic models enhance realism at the cost of interpretability. Behavior-based models focus on individual decision-making, demanding more computational resources. Data-driven approaches improve accuracy in nonlinear scenarios by adapting to evolving networks, using either traditional models or model-free machine learning techniques. We explore supervised and unsupervised learning methods, as well as reinforcement learning, which operates without predefined datasets. The application of graph neural networks (GNNs) is also discussed, highlighting their effectiveness in modeling propagation in complex networks. The paper under
Video world models can generate realistic futures from a single instruction, but they often fail to track the same physical points consistently across time. As a result, the generated videos appear plausible, yet lack the physical grounding required for reliable action execution, such as robot manipulation. We present GEM-4D, a geometry-grounded video world model that resolves this limitation by injecting dense 4D correspondence supervision distilled from a pretrained geometry foundation model into the video generative backbone during training. This supervision enables the model to jointly capture appearance and geometric structure while retaining a single-stream architecture with no additional inference cost. We further introduce an inverse dynamics module that converts correspondence-consistent video rollouts into executable robot trajectories, enabling direct deployment in both real-world and simulated manipulation. GEM-4D achieves state-of-the-art performance on both video prediction and geometric consistency across both simulation and realistic scenarios and improves real-world manipulation success from 61% to 81%. Additional results are available at https://gem-4d.github.io/.
Evaluating robot control policies is difficult: real-world testing is costly, and handcrafted simulators require manual effort to improve in realism and generality. We propose a world-model-based policy evaluation environment (WorldGym), an autoregressive, action-conditioned video generation model which serves as a proxy to real world environments. Policies are evaluated via Monte Carlo rollouts in the world model, with a vision-language model providing rewards. We evaluate a set of VLA-based real-robot policies in the world model using only initial frames from real robots, and show that policy success rates within the world model highly correlate with real-world success rates. Moreoever, we show that WorldGym is able to preserve relative policy rankings across different policy versions, sizes, and training checkpoints. Due to requiring only a single start frame as input, the world model further enables efficient evaluation of robot policies' generalization ability on novel tasks and environments. We find that modern VLA-based robot policies still struggle to distinguish object shapes and can become distracted by adversarial facades of objects. While generating highly realistic obj
We introduce Jais and Jais-chat, new state-of-the-art Arabic-centric foundation and instruction-tuned open generative large language models (LLMs). The models are based on the GPT-3 decoder-only architecture and are pretrained on a mixture of Arabic and English texts, including source code in various programming languages. With 13 billion parameters, they demonstrate better knowledge and reasoning capabilities in Arabic than any existing open Arabic and multilingual models by a sizable margin, based on extensive evaluation. Moreover, the models are competitive in English compared to English-centric open models of similar size, despite being trained on much less English data. We provide a detailed description of the training, the tuning, the safety alignment, and the evaluation of the models. We release two open versions of the model -- the foundation Jais model, and an instruction-tuned Jais-chat variant -- with the aim of promoting research on Arabic LLMs. Available at https://huggingface.co/inception-mbzuai/jais-13b-chat
Model merging via Mixture-of-Experts (MoE) has emerged as a scalable solution for consolidating multiple task-specific models into a unified sparse architecture, where each expert is derived from a model fine-tuned on a distinct task. While effective for multi-task integration, this paradigm introduces a critical yet underexplored challenge: how to attribute and protect the intellectual property (IP) of individual experts after merging. We propose RouteMark, a framework for IP protection in merged MoE models through the design of expert routing fingerprints. Our key insight is that task-specific experts exhibit stable and distinctive routing behaviors under probing inputs. To capture these patterns, we construct expert-level fingerprints using two complementary statistics: the Routing Score Fingerprint (RSF), quantifying the intensity of expert activation, and the Routing Preference Fingerprint (RPF), characterizing the input distribution that preferentially activates each expert. These fingerprints are reproducible, task-discriminative, and lightweight to construct. For attribution and tampering detection, we introduce a similarity-based matching algorithm that compares expert fin
Previous studies observed that finetuned models may be better base models than the vanilla pretrained model. Such a model, finetuned on some source dataset, may provide a better starting point for a new finetuning process on a desired target dataset. Here, we perform a systematic analysis of this intertraining scheme, over a wide range of English classification tasks. Surprisingly, our analysis suggests that the potential intertraining gain can be analyzed independently for the target dataset under consideration, and for a base model being considered as a starting point. This is in contrast to current perception that the alignment between the target dataset and the source dataset used to generate the base model is a major factor in determining intertraining success. We analyze different aspects that contribute to each. Furthermore, we leverage our analysis to propose a practical and efficient approach to determine if and how to select a base model in real-world settings. Last, we release an updating ranking of best models in the HuggingFace hub per architecture https://ibm.github.io/model-recycling/.
Below about 2.3 $μ$m, the nighttime emission of the Earth's atmosphere is dominated by non-thermal radiation from the mesosphere and thermosphere. As this airglow can even outshine scattered moonlight in the near-infrared regime, the understanding of the Earth's night-sky brightness requires good knowledge of the complex airglow emission spectrum and its variability. As airglow modelling is very challenging, the comprehensive characterisation of airglow emission requires large data sets of empirical data. For fixed locations, this can be best achieved by archived spectra of large astronomical telescopes with a wide wavelength coverage, high spectral resolving power, and good temporal sampling. Using 10 years of data from the X-shooter echelle spectrograph in the wavelength range from 0.3 to 2.5 $μ$m and additional data from the Ultraviolet and Visual Echelle Spectrograph at the Very Large Telescope at Cerro Paranal in Chile, we have succeeded to build a comprehensive spectroscopic airglow model for this low-latitude site under consideration of theoretical data from the HITRAN database for molecules and from different sources for atoms. The Paranal Airglow Line And Continuum Emissio
The advancement of Large Language Models (LLMs) has significantly transformed the field of natural language processing, although the focus on English-centric models has created a noticeable research gap for specific languages, including Vietnamese. To address this issue, this paper presents vi-mistral-x, an innovative Large Language Model designed expressly for the Vietnamese language. It utilizes a unique method of continual pre-training, based on the Mistral architecture, which incorporates grouped-query attention and sliding window attention techniques. This model, vi-Mistral-X, marks a significant step forward in improving the understanding and generation of the Vietnamese language. It introduces an additional phase of continual pre-training, specifically adapted for Vietnamese, enhancing the model's capability in understanding complex language nuances and generating accurate, context-aware Vietnamese text. Through comprehensive testing on various benchmarks, vi-mistral-x has shown to outperform existing Vietnamese LLMs in several key areas, including text classification, question answering, and text generation. Particularly, in the Vietnamese Multitask Language Understanding (
Diffusion models are highly effective at generating high-quality images but pose risks, such as the unintentional generation of NSFW (not safe for work) content. Although various techniques have been proposed to mitigate unwanted influences in diffusion models while preserving overall performance, achieving a balance between these goals remains challenging. In this work, we introduce EraseDiff, an algorithm designed to preserve the utility of the diffusion model on retained data while removing the unwanted information associated with the data to be forgotten. Our approach formulates this task as a constrained optimization problem using the value function, resulting in a natural first-order algorithm for solving the optimization problem. By altering the generative process to deviate away from the ground-truth denoising trajectory, we update parameters for preservation while controlling constraint reduction to ensure effective erasure, striking an optimal trade-off. Extensive experiments and thorough comparisons with state-of-the-art algorithms demonstrate that EraseDiff effectively preserves the model's utility, efficacy, and efficiency.
In the field of quantitative finance, volatility models, such as ARCH, GARCH, FIGARCH, SV, EWMA, play the key role in risk and portfolio management. Meanwhile, factor investing is more and more famous since mid of 20 century. CAPM, Fama French three factor model, Fama French five-factor model, MSCI Barra factor model are mentioned and developed during this period. In this paper, we will show why we need adjust group of factors by our MAXFLAT low-pass volatility model. All of our experiments are under China's CSI 300 and CSI 500 universe which represent China's large cap stocks and mid-small cap stocks. Our result shows adjust factors by MAXFLAT volatility model have better performance in both large cap and small cap universe than original factors or other risk adjust factors in China A share. Also the portfolio constructed by MAXFLAT risk adjust factors have continuous excess return and lower beta compare with benchmark index.
Is there a canonical way to think of agency beyond reward maximisation? In this paper, we show that any type of behaviour complying with physically sound assumptions about how macroscopic biological agents interact with the world canonically integrates exploration and exploitation in the sense of minimising risk and ambiguity about states of the world. This description, known as active inference, refines the free energy principle, a popular descriptive framework for action and perception originating in neuroscience. Active inference provides a normative Bayesian framework to simulate and model agency that is widely used in behavioural neuroscience, reinforcement learning (RL) and robotics. The usefulness of active inference for RL is three-fold. \emph{a}) Active inference provides a principled solution to the exploration-exploitation dilemma that usefully simulates biological agency. \emph{b}) It provides an explainable recipe to simulate behaviour, whence behaviour follows as an explainable mixture of exploration and exploitation under a generative world model, and all differences in behaviour are explicit in differences in world model. \emph{c}) This framework is universal in the
Significant advancements have been made in developing parametric models for digital humans, with various approaches concentrating on parts such as the human body, hand, or face. Nevertheless, connectors such as the neck have been overlooked in these models, with rich anatomical priors often unutilized. In this paper, we introduce HACK (Head-And-neCK), a novel parametric model for constructing the head and cervical region of digital humans. Our model seeks to disentangle the full spectrum of neck and larynx motions, facial expressions, and appearance variations, providing personalized and anatomically consistent controls, particularly for the neck regions. To build our HACK model, we acquire a comprehensive multi-modal dataset of the head and neck under various facial expressions. We employ a 3D ultrasound imaging scheme to extract the inner biomechanical structures, namely the precise 3D rotation information of the seven vertebrae of the cervical spine. We then adopt a multi-view photometric approach to capture the geometry and physically-based textures of diverse subjects, who exhibit a diverse range of static expressions as well as sequential head-and-neck movements. Using the mu
In this paper, we present the current position of the research project ML-Quadrat, which aims to extend the methodology, modeling language and tool support of ThingML - an open source modeling tool for IoT/CPS - to address Machine Learning needs for the IoT applications. Currently, ThingML offers a modeling language and tool support for modeling the components of the system, their communication interfaces as well as their behaviors. The latter is done through state machines. However, we argue that in many cases IoT/CPS services involve system components and physical processes, whose behaviors are not well understood in order to be modeled using state machines. Hence, quite often a data-driven approach that enables inference based on the observed data, e.g., using Machine Learning is preferred. To this aim, ML-Quadrat integrates the necessary Machine Learning concepts into ThingML both on the modeling level (syntax and semantics of the modeling language) and on the code generators level. We plan to support two target platforms for code generation regarding Stream Processing and Complex Event Processing, namely Apache SAMOA and Apama.
The complex and computationally expensive nature of landscape evolution models pose significant challenges in the inference and optimisation of unknown parameters. Bayesian inference provides a methodology for estimation and uncertainty quantification of unknown model parameters. In our previous work, we developed parallel tempering Bayeslands as a framework for parameter estimation and uncertainty quantification for the Badlands landscape evolution model. Parallel tempering Bayeslands features high-performance computing with dozens of processing cores running in parallel to enhance computational efficiency. Although we use parallel computing, the procedure remains computationally challenging since thousands of samples need to be drawn and evaluated. \textcolor{black}{In large-scale landscape and basin evolution problems, a single model evaluation can take from several minutes to hours, and in some instances, even days. Surrogate-assisted optimisation has been used for several computationally expensive engineering problems which motivate its use in optimisation and inference of complex geoscientific models.} The use of surrogate models can speed up parallel tempering Bayeslands by
We improve a recent mathematical model for cholera by adding a time delay that represents the time between the instant at which an individual becomes infected and the instant at which he begins to have symptoms of cholera disease. We prove that the delayed cholera model is biologically meaningful and analyze the local asymptotic stability of the equilibrium points for positive time delays. An optimal control problem is proposed and analyzed, where the goal is to obtain optimal treatment strategies, through quarantine, that minimize the number of infective individuals and the bacterial concentration, as well as treatment costs. Necessary optimality conditions are applied to the delayed optimal control problem, with a $L^1$ type cost functional. We show that the delayed cholera model fits better the cholera outbreak that occurred in the Department of Artibonite -- Haiti, from 1 November 2010 to 1 May 2011, than the non-delayed model. Considering the data of the cholera outbreak in Haiti, we solve numerically the delayed optimal control problem and propose solutions for the outbreak control and eradication.
Human traffic along roads can be a major vector for infectious diseases and invasive species. Though most road traffic is local, a small number of long-distance trips can suffice to move an invasion or disease front forward. Therefore, understanding how many agents travel over long distances and which routes they choose is key to successful management of diseases and invasions. Stochastic gravity models have been used to estimate the distribution of trips between origins and destinations of agents. However, in large-scale systems it is hard to collect the data required to fit these models, as the number of long-distance travellers is small, and origins and destinations can have multiple access points. Therefore, gravity models often provide only relative measures of the agent flow. Furthermore, gravity models yield no insights into which roads agents use. We resolve these issues by combining a stochastic gravity model with a stochastic route choice model. Our hybrid model can be fitted to survey data collected at roads that are used by many long-distance travellers. This decreases the sampling effort, allows us to obtain absolute predictions of both vector pressure and pathways, an
We present HunyuanImage 3.0, a native multimodal model that unifies multimodal understanding and generation within an autoregressive framework, with its image generation module publicly available. The achievement of HunyuanImage 3.0 relies on several key components, including meticulous data curation, advanced architecture design, a native Chain-of-Thoughts schema, progressive model pre-training, aggressive model post-training, and an efficient infrastructure that enables large-scale training and inference. With these advancements, we successfully trained a Mixture-of-Experts (MoE) model comprising over 80 billion parameters in total, with 13 billion parameters activated per token during inference, making it the largest and most powerful open-source image generative model to date. We conducted extensive experiments and the results of automatic and human evaluation of text-image alignment and visual quality demonstrate that HunyuanImage 3.0 rivals previous state-of-the-art models. By releasing the code and weights of HunyuanImage 3.0, we aim to enable the community to explore new ideas with a state-of-the-art foundation model, fostering a dynamic and vibrant multimodal ecosystem. Al
Training effective software engineering agents requires large volumes of task-specific trajectories, incurring substantial data construction costs. Inspired by the "Less-Is-More" hypothesis in mathematical reasoning, we investigate its extension to agentic scenarios and propose an end-to-end training framework that achieves superior agentic capabilities with fewer but higher-quality training trajectories. This is achieved via STITCH (Sliding-memory Trajectory Inference and Task Chunking Heuristic), a coarse-to-fine mechanism that filters low-value noise and retains decision-critical tokens to maximize training signal quality. We conduct experiments across multiple agent frameworks (e.g., mini-SWE-agent, MSWE-agent), model scales (30B to 355B), and multilingual settings (Python, Java, and ArkTS). On SWE-bench Verified, models trained with STITCH achieve up to 63.16% relative improvement over base models. On Multi-SWE-bench (Java), MiniMax-M2.5-STITCH achieves 43.75% with our CodeArts Agent scaffold (+16.67%). On HarmonyOS (ArkTS), GLM-4.7-STITCH improves the compilation pass rate to 61.31% (+43.34%) with less than 1K training trajectories. Our results confirm that the "Less-Is-More"
We present Mach-Mind-4-Flash, a 35B-parameter Mixture-of-Experts (MoE) agentic model with 3B activated parameters. Through post-training optimization alone without scaling pre-training compute, the model achieves performance on par with or surpassing that of 100B-parameter-class models. By introducing scalable agentic interaction environments for large-scale reinforcement learning, the model attains significant performance gains on real-world application tasks. Our pipeline comprises three stages: (1) a unified RL/OPD training infrastructure with dynamic multi-teacher scheduling and operator-level acceleration, delivering 17\% end-to-end training speedup; (2) multiple domain-specific RL experts trained in parallel across Reasoning, General, and Agent tracks, then fused into a single generalist via Multi-Teacher On-Policy Distillation (MOPD) -- a routed reverse-KL objective that eliminates the see-saw degradation of mixed-reward RL; (3) Hybrid Median-length Policy Optimization (HMPO), a single-stage token-efficiency method that compresses reasoning chains by 19--46\% with $\le$0.7 percentage-point accuracy loss. Mach-Mind-4-Flash scores 92.70 on AIME'26, 82.82 on IFBench, 80.74 on B