Search

Efficient multiple setpoint tracking can enable advanced biotechnological applications, such as maintaining desired population levels in co-cultures for optimal metabolic division of labor. In this study, we employ reinforcement learning as a control method for population setpoint tracking in co-cultures, focusing on policy-gradient techniques where the control policy is parameterized by neural networks. However, achieving accurate tracking across multiple setpoints is a significant challenge in reinforcement learning, as the agent must effectively balance the contributions of various setpoints to maximize the expected system performance. Traditional return functions, such as those based on a quadratic cost, often yield suboptimal performance due to their inability to efficiently guide the agent toward the simultaneous satisfaction of all setpoints. To overcome this, we propose a novel return function that rewards the simultaneous satisfaction of multiple setpoints and diminishes overall reward gains otherwise, accounting for both stage and terminal system performance. This return function includes parameters to fine-tune the desired smoothness and steepness of the learning process. We demonstrate our approach considering an $\textit{Escherichia coli}$ co-culture in a chemostat with optogenetic control over amino acid synthesis pathways, leveraging auxotrophies to modulate growth.

Large Language Models (LLMs) have attracted significant attention due to their human-like language understanding and generation capabilities, as well as their applicability across various domains. These models, characterized by their massive scale and extensive training data, continue to push the boundaries of what is possible in natural language processing. The Llama 3 series, for instance, exemplifies this trend with its flagship model boasting 405 billion parameters trained on 15.6 trillion tokens. The immense computational demands associated with training such models have spurred ongoing research into optimizing the efficiency of the training process, particularly through the use of lower-precision formats. NVIDIA's H100 GPU, which introduces support for FP8 in addition to the more conventional FP16 and BF16 formats, has emerged as a focal point in this optimization effort. Preliminary studies suggest that FP8 could offer substantial reductions in training time without sacrificing model performance when compared to BF16, making it a promising candidate for large-scale model training. However, the broader implications of adopting FP8, particularly in terms of training stability and downstream task performance, have yet to be fully understood. In this study, we delve into the practical trade-offs involved in adopting FP8 over BF16 for training LLMs., Comment: 2 pages,extended abstract

In large language model (LLM) training, several parallelization strategies, including Tensor Parallelism (TP), Pipeline Parallelism (PP), Data Parallelism (DP), as well as Sequence Parallelism (SP) and Context Parallelism (CP), are employed to distribute model parameters, activations, and optimizer states across devices. Identifying the optimal parallelization configuration for each environment while avoiding GPU memory overflow remains a challenging task. In this study, we provide precise formulas to estimate the memory consumed by parameters, gradients, optimizer states, and activations for 4D parallel training (DP, TP, PP, CP) in the Llama architecture. We conducted 454 experiments on A100 and H100 GPUs, incorporating often neglected factors such as temporary buffers and memory fragmentation into our analysis. Results indicate that when the estimated memory usage is below 80\% of the available GPU memory, the training never encounters out-of-memory errors. This simple yet effective formula allows us to identify parallelization configurations that could lead to memory overflow in advance, significantly reducing the configuration search space. Additionally, through a comprehensive exploration of optimal configurations in 4D parallelism, our analysis of the 454 experimental results provides empirical insights into optimal 4D parallelism configurations.

We show that variational learning can significantly improve the accuracy and calibration of Low-Rank Adaptation (LoRA) without a substantial increase in the cost. We replace AdamW by the Improved Variational Online Newton (IVON) algorithm to finetune large language models. For Llama-2 with 7 billion parameters, IVON improves the accuracy over AdamW by 2.8% and expected calibration error by 4.6%. The accuracy is also better than the other Bayesian alternatives, yet the cost is lower and the implementation is easier. Our work provides additional evidence for the effectiveness of IVON for large language models. The code is available at https://github.com/team-approx-bayes/ivon-lora., Comment: Published at 38th Workshop on Fine-Tuning in Machine Learning (NeurIPS 2024). Code available at https://github.com/team-approx-bayes/ivon-lora. In version 2 we fixed a typo in the equation of prior in section 2

The Privacy Preserving Federated Learning Document VQA (PFL-DocVQA) competition challenged the community to develop provably private and communication-efficient solutions in a federated setting for a real-life use case: invoice processing. The competition introduced a dataset of real invoice documents, along with associated questions and answers requiring information extraction and reasoning over the document images. Thereby, it brings together researchers and expertise from the document analysis, privacy, and federated learning communities. Participants fine-tuned a pre-trained, state-of-the-art Document Visual Question Answering model provided by the organizers for this new domain, mimicking a typical federated invoice processing setup. The base model is a multi-modal generative language model, and sensitive information could be exposed through either the visual or textual input modality. Participants proposed elegant solutions to reduce communication costs while maintaining a minimum utility threshold in track 1 and to protect all information from each document provider using differential privacy in track 2. The competition served as a new testbed for developing and testing private federated learning methods, simultaneously raising awareness about privacy within the document image analysis and recognition community. Ultimately, the competition analysis provides best practices and recommendations for successfully running privacy-focused federated learning challenges in the future., Comment: 27 pages, 6 figures

Local learning, which trains a network through layer-wise local targets and losses, has been studied as an alternative to backpropagation (BP) in neural computation. However, its algorithms often become more complex or require additional hyperparameters because of the locality, making it challenging to identify desirable settings in which the algorithm progresses in a stable manner. To provide theoretical and quantitative insights, we introduce the maximal update parameterization ($\mu$P) in the infinite-width limit for two representative designs of local targets: predictive coding (PC) and target propagation (TP). We verified that $\mu$P enables hyperparameter transfer across models of different widths. Furthermore, our analysis revealed unique and intriguing properties of $\mu$P that are not present in conventional BP. By analyzing deep linear networks, we found that PC's gradients interpolate between first-order and Gauss-Newton-like gradients, depending on the parameterization. We demonstrate that, in specific standard settings, PC in the infinite-width limit behaves more similarly to the first-order gradient. For TP, even with the standard scaling of the last layer, which differs from classical $\mu$P, its local loss optimization favors the feature learning regime over the kernel regime.

PC-Gym is an open-source tool designed to facilitate the development and evaluation of reinforcement learning (RL) algorithms for chemical process control problems. It provides a suite of environments that model a range of chemical processes, incorporating nonlinear dynamics, process disturbances, and constraints. Key features include flexible constraint handling mechanisms, customizable disturbance generation, and modular reward function design. The framework enables benchmarking state-of-the-art RL algorithms against a nonlinear Model Predictive Control (NMPC) oracle across various process control scenarios. Case studies demonstrate PC-Gym's effectiveness in evaluating RL approaches for the control of various chemical engineering systems such as a continuously stirred tank reactor, multistage extraction process, and crystallization reactor. The framework's ability to incorporate realistic disturbances and constraints allows for robust testing of control strategies. Results highlight the performance gaps between RL algorithms and NMPC oracles, demonstrating the utility of PC-Gym for algorithm benchmarking and suggesting areas for improvement in RL-based process control. By offering a standardized platform for developing and assessing RL-based control strategies, PC-Gym aims to accelerate research at the intersection of machine learning and process systems engineering. It bridges the gap between theoretical advancements in RL and practical applications in industrial process control, providing researchers and practitioners with a valuable tool for exploring data-driven control solutions for complex chemical processes.

Microkinetic models are key for evaluating industrial processes' efficiency and chemicals' environmental impact. Manual construction of these models is difficult and time-consuming, prompting a shift to automated methods. This study introduces SiMBA (Simplest Mechanism Builder Algorithm), a novel approach for generating microkinetic models from kinetic data. SiMBA operates through four phases: mechanism generation, mechanism translation, parameter estimation, and model comparison. Our approach systematically proposes reaction mechanisms, using matrix representations and a parallelized backtracking algorithm to manage complexity. These mechanisms are then translated into microkinetic models represented by ordinary differential equations, and optimized to fit available data. Models are compared using information criteria to balance accuracy and complexity, iterating until convergence to an optimal model is reached. Case studies on an aldol condensation reaction, and the dehydration of fructose demonstrate SiMBA's effectiveness in distilling complex kinetic behaviors into simple yet accurate models. While SiMBA predicts intermediates correctly for all case studies, it does not chemically identify intermediates, requiring expert input for complex systems. Despite this, SiMBA significantly enhances mechanistic exploration, offering a robust initial mechanism that accelerates the development and modeling of chemical processes.

Inventory control in modern supply chains has attracted significant attention due to the increasing number of disruptive shocks and the challenges posed by complex dynamics, uncertainties, and limited collaboration. Traditional methods, which often rely on static parameters, struggle to adapt to changing environments. This paper proposes a Multi-Agent Reinforcement Learning (MARL) framework with Graph Neural Networks (GNNs) for state representation to address these limitations. Our approach redefines the action space by parameterizing heuristic inventory control policies, making it adaptive as the parameters dynamically adjust based on system conditions. By leveraging the inherent graph structure of supply chains, our framework enables agents to learn the system's topology, and we employ a centralized learning, decentralized execution scheme that allows agents to learn collaboratively while overcoming information-sharing constraints. Additionally, we incorporate global mean pooling and regularization techniques to enhance performance. We test the capabilities of our proposed approach on four different supply chain configurations and conduct a sensitivity analysis. This work paves the way for utilizing MARL-GNN frameworks to improve inventory management in complex, decentralized supply chain environments.

The gravitational physics landscape is evolving rapidly, driven by our ability to study strong-field regions, in particular black holes. Black Holes Inside and Out gathered world experts to discuss the status of the field and prospects ahead. We hope that the ideas and perspectives are a source of inspiration. Structure: Black Hole Evaporation - 50 Years by William Unruh The Stability Problem for Extremal Black Holes by Mihalis Dafermos The Entropy of Black Holes by Robert M. Wald The Non-linear Regime of Gravity by Luis Lehner Black Holes Galore in D > 4 by Roberto Emparan Same as Ever: Looking for (In)variants in the Black Holes Landscape by Carlos A. R. Herdeiro Black Holes, Cauchy Horizons, and Mass Inflation by Matt Visser The Backreaction Problem for Black Holes in Semiclassical Gravity by Adrian del Rio Black Holes Beyond General Relativity by Enrico Barausse and Jutta Kunz Black Holes as Laboratories: Searching for Ultralight Fields by Richard Brito Primordial Black Holes from Inflation by Misao Sasaki Tests of General Relativity with Future Detectors by Emanuele Berti Black Holes as Laboratories: Tests of General Relativity by Ruth Gregory and Samaya Nissanke Simulating Black Hole Imposters by Frans Pretorius Black Hole Spectroscopy: Status Report by Gregorio Carullo VLBI as a Precision Strong Gravity Instrument by Paul Tiede Testing the nature of compact objects and the black hole paradigm by Mariafelicia De Laurentis and Paolo Pani Some Thoughts about Black Holes in Asymptotic Safety by Alessia Platania Black Hole Evaporation in Loop Quantum Gravity by Abhay Ashtekar How the Black Hole Puzzles are Resolved in String Theory by Samir D. Mathur Quantum Black Holes: From Regularization to Information Paradoxes by Niayesh Afshordi and Stefano Liberati, Comment: 221 pages, 21 contributions

We introduce a model that can accurately simulate radiation from undulator sources for ray tracing applications. It incorporates several key effects relevant to 4$^\text{th}$ generation synchrotron sources, such as electron emittance, energy spread, and diffraction-limited beam size. This code has been developed as part of SHADOW4, the latest version of the widely used SHADOW X-ray optics ray tracing program. The approach relies on calculating the field distribution in the far field, which determines the ray divergences. The integration of existing models for electron energy spread is also addressed. Rays sampled at the source follow a size distribution derived by backpropagating the far field radiation. These models are detailed, and several examples are provided., Comment: 12 page, 12 figures

Vision Language Models (VLMs) often struggle with culture-specific knowledge, particularly in languages other than English and in underrepresented cultural contexts. To evaluate their understanding of such knowledge, we introduce WorldCuisines, a massive-scale benchmark for multilingual and multicultural, visually grounded language understanding. This benchmark includes a visual question answering (VQA) dataset with text-image pairs across 30 languages and dialects, spanning 9 language families and featuring over 1 million data points, making it the largest multicultural VQA benchmark to date. It includes tasks for identifying dish names and their origins. We provide evaluation datasets in two sizes (12k and 60k instances) alongside a training dataset (1 million instances). Our findings show that while VLMs perform better with correct location context, they struggle with adversarial contexts and predicting specific regional cuisines and languages. To support future research, we release a knowledge base with annotated food entries and images along with the VQA data., Comment: Preprint

In the modern world, the development of Artificial Intelligence (AI) has contributed to improvements in various areas, including automation, computer vision, fraud detection, and more. AI can be leveraged to enhance the efficiency of Autonomous Smart Traffic Management (ASTM) systems and reduce traffic congestion rates. This paper presents an Autonomous Smart Traffic Management (STM) system that uses AI to improve traffic flow rates. The system employs the YOLO V5 Convolutional Neural Network to detect vehicles in traffic management images. Additionally, it predicts the number of vehicles for the next 12 hours using a Recurrent Neural Network with Long Short-Term Memory (RNN-LSTM). The Smart Traffic Management Cycle Length Analysis manages the traffic cycle length based on these vehicle predictions, aided by AI. From the results of the RNN-LSTM model for predicting vehicle numbers over the next 12 hours, we observe that the model predicts traffic with a Mean Squared Error (MSE) of 4.521 vehicles and a Root Mean Squared Error (RMSE) of 2.232 vehicles. After simulating the STM system in the CARLA simulation environment, we found that the Traffic Management Congestion Flow Rate with ASTM (21 vehicles per minute) is 50\% higher than the rate without STM (around 15 vehicles per minute). Additionally, the Traffic Management Vehicle Pass Delay with STM (5 seconds per vehicle) is 70\% lower than without STM (around 12 seconds per vehicle). These results demonstrate that the STM system using AI can increase traffic flow by 50\% and reduce vehicle pass delays by 70\%., Comment: In process to IEEE Intelligent Vehicle Symposium 2025

Point cloud completion networks are conventionally trained to minimize the disparities between the completed point cloud and the ground-truth counterpart. However, an incomplete object-level point cloud can have multiple valid completion solutions when it is examined in isolation. This one-to-many mapping issue can cause contradictory supervision signals to the network because the loss function may produce different values for identical input-output pairs of the network. In many cases, this issue could adversely affect the network optimization process. In this work, we propose to enhance the conventional learning objective using a novel completion consistency loss to mitigate the one-to-many mapping problem. Specifically, the proposed consistency loss ensure that a point cloud completion network generates a coherent completion solution for incomplete objects originating from the same source point cloud. Experimental results across multiple well-established datasets and benchmarks demonstrated the proposed completion consistency loss have excellent capability to enhance the completion performance of various existing networks without any modification to the design of the networks. The proposed consistency loss enhances the performance of the point completion network without affecting the inference speed, thereby increasing the accuracy of point cloud completion. Notably, a state-of-the-art point completion network trained with the proposed consistency loss can achieve state-of-the-art accuracy on the challenging new MVP dataset. The code and result of experiment various point completion models using proposed consistency loss will be available at: https://github.com/kaist-avelab/ConsistencyLoss ., Comment: First version of Paper "Enhancing Performance of Point Cloud Completion Networks with Consistency Loss" by Kevin Tirta Wijaya and Christofel Rio Goenawan. In process submission to Neurocomputing Journal 2024

Today, we are open-sourcing our core speech recognition and diarization models for non-commercial use. We are releasing both a full production pipeline for developers as well as pared-down research models for experimentation. Rev hopes that these releases will spur research and innovation in the fast-moving domain of voice technology. The speech recognition models released today outperform all existing open source speech recognition models across a variety of long-form speech recognition domains.

In a previous work (Hern\'andez-Bernal et al. 2021) we performed an observational analysis of the Arsia Mons Elongated Cloud (AMEC), which stands out due to its impressive size and shape, quick dynamics, and the fact that it happens during the martian dusty season. Observations show that its morphology can be split in a head, on the western slope of the volcano of around 120 km in diameter; and a tail, that expands to the west reaching more than 1000 km in length, making the AMEC the longest orographic cloud observed so far in the solar system. In this work we run the LMD (Laboratoire de M\'et\'eorologie Dynamique) Mesoscale Model to gain insight into the physics of the AMEC. We note that it is coincident in terms of local time and seasonality with the fastest winds on the summit of Arsia Mons. A downslope windstorm on the western slope is followed by a hydraulic-like jump triggering a strong vertical updraft that propagates upwards in the atmosphere, causing a drop in temperatures of down to 30K at 40-50 km in altitude, spatially and temporarily coincident with the observed head of the AMEC. However the model does not reproduce the microphysics of this cloud: the optical depth is too low and the expansion of the tail does not happen in the model. The observed diurnal cycle is correctly captured by the model for the head of the cloud. This work raises new questions that will guide future observations of the AMEC.

The Visual Monitoring Camera (VMC) is a small imaging instrument onboard Mars Express with a field of view of ~40x30 degrees. The camera was initially intended to provide visual confirmation of the separation of the Beagle 2 lander and has similar technical specifications to a typical webcam of the 2000s. In 2007, a few years after the end of its original mission, VMC was turned on again to obtain full-disk images of Mars to be used for outreach purposes. As VMC obtained more images, the scientific potential of the camera became evident, and in 2018 the camera was given an upgraded status of a new scientific instrument, with science goals in the field of Martian atmosphere meteorology. The wide Field of View of the camera combined with the orbit of Mars Express enable the acquisition of full-disk images of the planet showing different local times, which for a long time has been rare among orbital missions around Mars. The small data volume of images also allows videos that show the atmospheric dynamics of dust and cloud systems to be obtained. This paper is intended to be the new reference paper for VMC as a scientific instrument, and thus provides an overview of the updated procedures to plan, command and execute science observations of the Martian atmosphere. These observations produce valuable science data that is calibrated and distributed to the community for scientific use.

We present WOFRY (Wave Optics FRamework in pYthon), a specialized toolbox designed for wave optics modeling, with particular emphasis on partial coherence. This package is tailored to assist synchrotron scientists and engineers in the prototyping and modeling of X-ray sources and optics. WOFRY offers functionalities for generating and propagating 1D and 2D wavefronts, along with various tools for interacting with different optical elements. The software developed is available in the OASYS suite., Comment: 13 pages, 8 figures

In positron emission tomography acquisition (PET), sensitivity along a line of response can vary due to crystal geometrical arrangements in the scanner and/or detector inefficiencies, leading to severe artefacts in the reconstructed image. To mitigate these effects, data must be corrected by a set of normalization coefficients applied to each line of response. The J-PET Modular scanner is a PET device made of 50 cm long plastic strips arranged axially, currently in operation at the Jagiellonian University in Krak\'ow (Poland). We have implemented a normalization method for the large field-of-view plastic-based J-PET Modular scanner using the component-based approach. We estimated the geometric normalization factors for the J-PET Modular scanner using Monte Carlo simulations. We also analysed the effects of variations in detection efficiency. A dedicated cylindrical phantom was simulated to investigate the impact of various factors on image quality. The image quality was quantified in terms of radial and axial uniformity metrics, and the standard deviation to mean intensity ratio, determined for a set of image slices. Without normalization, reconstructions of a uniform cylinder exhibit artefacts. These artefacts were satisfactorily compensated using the normalization factors. Applying geometrical corrections lowers the non-uniformity of the image expressed as a standard deviation-to-mean ratio to a range between 5.5 % to 8.5 %. Computationally, the technique is straightforward to parallelize, making it time-efficient. Preliminary estimates suggest that the method is appropriate for use with long axial field-of-view scanners, such as the total-body J-PET, currently under development at the Jagiellonian University., Comment: 36 pages, 16 figures, 4 tables

The positronium, a bound state of electron and positron is a unique system to perform highly precise tests, due to no hadronic background and precise Quantum Electrodynamics (QED) predictions. Being a system of lepton and antilepton, its properties are precisely described by QED in the Standard Model (SM). The final events topology can be simulated using Monte Carlo techniques. The J-PET detector is a multi-purpose, large acceptance system that is very well-suitable to the studies of positronium decay due to its excellent angular (1$^o$) and timing resolutions. \\ We present preliminary results on the feasibility of searching for Dark Matter (DM) candidates in the decay o-Ps $\rightarrow$ invisible with the J-PET, which is well suited for the detection of positronium decay products. Toy Monte Carlo simulations have been prepared to incorporate DM decay models to the o-Ps decay expectations in order to assess the detector capabilities to search for such an elusive component of our Universe.

Mobile health has the potential to revolutionize health care delivery and patient engagement. In this work, we discuss how integrating Artificial Intelligence into digital health applications-focused on supply chain, patient management, and capacity building, among other use cases-can improve the health system and public health performance. We present an Artificial Intelligence and Reinforcement Learning platform that allows the delivery of adaptive interventions whose impact can be optimized through experimentation and real-time monitoring. The system can integrate multiple data sources and digital health applications. The flexibility of this platform to connect to various mobile health applications and digital devices and send personalized recommendations based on past data and predictions can significantly improve the impact of digital tools on health system outcomes. The potential for resource-poor settings, where the impact of this approach on health outcomes could be more decisive, is discussed specifically. This framework is, however, similarly applicable to improving efficiency in health systems where scarcity is not an issue., Comment: This article has been accepted for publication in Health Systems & Reform, published by Taylor & Francis

Throughout the history of computer vision, while research has explored the integration of images (visual) and point clouds (geometric), many advancements in image and 3D object recognition have tended to process these modalities separately. We aim to bridge this divide by integrating images and point clouds on a unified transformer model. This approach integrates the modality-specific properties of images and point clouds and achieves fundamental downstream tasks in image and 3D object recognition on a unified transformer model by learning visual-geometric representations. In this work, we introduce Formula-Supervised Visual-Geometric Pre-training (FSVGP), a novel synthetic pre-training method that automatically generates aligned synthetic images and point clouds from mathematical formulas. Through cross-modality supervision, we enable supervised pre-training between visual and geometric modalities. FSVGP also reduces reliance on real data collection, cross-modality alignment, and human annotation. Our experimental results show that FSVGP pre-trains more effectively than VisualAtom and PC-FractalDB across six tasks: image and 3D object classification, detection, and segmentation. These achievements demonstrate FSVGP's superior generalization in image and 3D object recognition and underscore the potential of synthetic pre-training in visual-geometric representation learning. Our project website is available at https://ryosuke-yamada.github.io/fdsl-fsvgp/., Comment: Accepted to ECCV2024

Positron Annihilation Lifetime Spectroscopy (PALS) is a well-established non-destructive technique used for nanostructural characterization of porous materials. It is based on the annihilation of a positron and an electron. Mean positron lifetime in the material depends on the free voids size and molecular environment, allowing the study of porosity and structural transitions in the nanometer scale. We have developed a novel method enabling spatially resolved PALS, thus providing tomography of nanostructural characterization of an extended object. Correlating space (position) and structural (lifetime) information brings new insight in materials studies, especially in the characterization of the purity and pore distribution. For the first time, a porosity image using stationary positron sources for the simultaneous measurement of the porous polymers XAD4, silica aerogel powder IC3100, and polyvinyl toluene scintillator PVT by the J-PET tomograph is demonstrated

Multi fidelity Bayesian optimization (MFBO) leverages experimental and or computational data of varying quality and resource cost to optimize towards desired maxima cost effectively. This approach is particularly attractive for chemical discovery due to MFBO's ability to integrate diverse data sources. Here, we investigate the application of MFBO to accelerate the identification of promising molecules or materials. We specifically analyze the conditions under which lower fidelity data can enhance performance compared to single-fidelity problem formulations. We address two key challenges, selecting the optimal acquisition function, understanding the impact of cost, and data fidelity correlation. We then discuss how to assess the effectiveness of MFBO for chemical discovery.

This work presents a detailed case study on using Generative AI (GenAI) to develop AI surrogates for simulation models in fusion energy research. The scope includes the methodology, implementation, and results of using GenAI to assist in model development and optimization, comparing these results with previous manually developed models.

During coating processes, dust deposition can lead to an uneven thickness in the resulting film, posing significant problems in industrial processes. Our study explores the effects of solid defects using a vertical cylindrical fiber deposited on a silicone oil film coating a horizontal solid substrate. We use a hyperspectral camera to measure the film thickness by interferometry in the vicinity of the defect. As predicted and observed in many studies in various geometries, a circular groove appears around the fiber because of the capillary suction induced by the meniscus that grows at the bottom of the fiber. We measure the evolution of the thickness of the film at the groove over time. The thickness decreases before increasing again leading to the healing of the perturbation at long time. We propose that healing is due to the arrest of the suction when the meniscus reaches its equilibrium shape. By combining geometric analysis with the thin film equation, we have developed scaling laws that predict both the minimum thickness of the groove, that we call the time-minimum thickness, and the time required to reach this minimum. If the time-minimum thickness reaches the thickness at which intermolecular forces begin to play a role prior to healing, the thickness of the groove will stop decreasing and saturate due to the competition between drainage and repulsive intermolecular forces. Based on the previous scaling law, we developed a scaling law accounting for the critical initial thickness of the film below which the intermolecular repulsion will start to have an effect, which is in good agreement with our experiments. These results thus offer valuable insights into predicting and preventing defects in coating processes, thereby improving the quality and reliability of coated products in various industries.

Common measures of accuracy used to assess the performance of automatic speech recognition (ASR) systems, as well as human transcribers, conflate multiple sources of error. Stylistic differences, such as verbatim vs non-verbatim, can play a significant role in ASR performance evaluation when differences exist between training and test datasets. The problem is compounded for speech from underrepresented varieties, where the speech to orthography mapping is not as standardized. We categorize the kinds of stylistic differences between 6 transcription versions, 4 human- and 2 ASR-produced, of 10 hours of African American English (AAE) speech. Focusing on verbatim features and AAE morphosyntactic features, we investigate the interactions of these categories with how well transcripts can be compared via word error rate (WER). The results, and overall analysis, help clarify how ASR outputs are a function of the decisions made by the training data's human transcribers., Comment: Published in Interspeech 2024 Proceedings, 5 pages excluding references, 5 figures

In this work, we investigate the understudied effect of the training data used for image super-resolution (SR). Most commonly, novel SR methods are developed and benchmarked on common training datasets such as DIV2K and DF2K. However, we investigate and rethink the training data from the perspectives of diversity and quality, {thereby addressing the question of ``How important is SR training for SR models?''}. To this end, we propose an automated image evaluation pipeline. With this, we stratify existing high-resolution image datasets and larger-scale image datasets such as ImageNet and PASS to compare their performances. We find that datasets with (i) low compression artifacts, (ii) high within-image diversity as judged by the number of different objects, and (iii) a large number of images from ImageNet or PASS all positively affect SR performance. We hope that the proposed simple-yet-effective dataset curation pipeline will inform the construction of SR datasets in the future and yield overall better models., Comment: Accepted to ECCV2024

This work proposes a control-informed reinforcement learning (CIRL) framework that integrates proportional-integral-derivative (PID) control components into the architecture of deep reinforcement learning (RL) policies. The proposed approach augments deep RL agents with a PID controller layer, incorporating prior knowledge from control theory into the learning process. CIRL improves performance and robustness by combining the best of both worlds: the disturbance-rejection and setpoint-tracking capabilities of PID control and the nonlinear modeling capacity of deep RL. Simulation studies conducted on a continuously stirred tank reactor system demonstrate the improved performance of CIRL compared to both conventional model-free deep RL and static PID controllers. CIRL exhibits better setpoint-tracking ability, particularly when generalizing to trajectories outside the training distribution, suggesting enhanced generalization capabilities. Furthermore, the embedded prior control knowledge within the CIRL policy improves its robustness to unobserved system disturbances. The control-informed RL framework combines the strengths of classical control and reinforcement learning to develop sample-efficient and robust deep reinforcement learning algorithms, with potential applications in complex industrial systems.

We have performed a parametrized post-Newtonian (PPN) test of a black hole spacetime using observational data of the star S0-2/S2 orbiting the massive black hole at our galactic center Sgr A*. After introducing our PPN model of black hole spacetime, we report the result of $\chi^2$ fitting of the PPN model with the observational data. A new finding through our PPN model is a detectability of the gravitational lens effect on the null geodesics connecting S0-2 and observer under the present observational uncertainties, if a PPN parameter is about one order larger than the value for general relativity case. On the other hand, the effect of black hole spin on the S0-2's motion is not detectable. Thus our present PPN test is performed with spherically symmetric vacuum black hole spacetime. The resultant value of the PPN parameter, which corresponds to the minimum $\chi^2$, implies that the gravitational field of Sgr A* is not of Schwarzschild metric or that there exists a sufficient amount of dark matters around Sgr A* to be detected by present telescopes. However, the difference between the minimum $\chi^2$ and the $\chi^2$ of Schwarzschild case is not large enough to ensure a statistical significance of non-Schwarzschild result. A more precise statistical analysis than $\chi^2$ statistics is necessary to extract a statistically significant information of the gravitational field of Sgr A* from present observational data. We will report a result by a Bayesian analysis in next paper., Comment: 43 pages, 4 figures, Accepted for Publication in Prog.Theor.Exp.Phys / Typos are corrected

We prove that any tensor product factorization with a commutative factor of a modular group algebra over a prime field comes from a direct product decomposition of the group basis. This extends previous work by Carlson and Kov\'acs for the commutative case and answers a question of them in some cases.

This paper introduces a reinforcement learning (RL) platform that enhances end-to-end user journeys in healthcare digital tools through personalization. We explore a case study with SwipeRx, the most popular all-in-one app for pharmacists in Southeast Asia, demonstrating how the platform can be used to personalize and adapt user experiences. Our RL framework is tested through a series of experiments with product recommendations tailored to each pharmacy based on real-time information on their purchasing history and in-app engagement, showing a significant increase in basket size. By integrating adaptive interventions into existing mobile health solutions and enriching user journeys, our platform offers a scalable solution to improve pharmaceutical supply chain management, health worker capacity building, and clinical decision and patient care, ultimately contributing to better healthcare outcomes., Comment: Presented at the Third Workshop on End-to-End Customer Journey Optimization at KDD 2024 (KDD CJ Workshop '24), August 26, Barcelona, Spain

Pharmacies are critical in healthcare systems, particularly in low- and middle-income countries. Procuring pharmacists with the right behavioral interventions or nudges can enhance their skills, public health awareness, and pharmacy inventory management, ensuring access to essential medicines that ultimately benefit their patients. We introduce a reinforcement learning operational system to deliver personalized behavioral interventions through mobile health applications. We illustrate its potential by discussing a series of initial experiments run with SwipeRx, an all-in-one app for pharmacists, including B2B e-commerce, in Indonesia. The proposed method has broader applications extending beyond pharmacy operations to optimize healthcare delivery., Comment: Presented at The First Workshop on AI Behavioral Science (AIBS'24) at KDD 2024, August 25, Barcelona, Spain

By providing evidence-based clinical decision support, digital tools and electronic health records can revolutionize patient management, especially in resource-poor settings where fewer health workers are available and often need more training. When these tools are integrated with AI, they can offer personalized support and adaptive interventions, effectively connecting community health workers (CHWs) and healthcare facilities. The CHARM (Community Health Access & Resource Management) app is an AI-native mobile app for CHWs. Developed through a joint partnership of Causal Foundry (CF) and mothers2mothers (m2m), CHARM empowers CHWs, mainly local women, by streamlining case management, enhancing learning, and improving communication. This paper details CHARM's development, integration, and upcoming reinforcement learning-based adaptive interventions, all aimed at enhancing health worker engagement, efficiency, and patient outcomes, thereby enhancing CHWs' capabilities and community health., Comment: Presented at the 7th epiDAMIK ACM SIGKDD International Workshop on Epidemiology meets Data Mining and Knowledge Discovery, August 26, 2024, Barcelona, Spain

Pre-training and transfer learning are an important building block of current computer vision systems. While pre-training is usually performed on large real-world image datasets, in this paper we ask whether this is truly necessary. To this end, we search for a minimal, purely synthetic pre-training dataset that allows us to achieve performance similar to the 1 million images of ImageNet-1k. We construct such a dataset from a single fractal with perturbations. With this, we contribute three main findings. (i) We show that pre-training is effective even with minimal synthetic images, with performance on par with large-scale pre-training datasets like ImageNet-1k for full fine-tuning. (ii) We investigate the single parameter with which we construct artificial categories for our dataset. We find that while the shape differences can be indistinguishable to humans, they are crucial for obtaining strong performances. (iii) Finally, we investigate the minimal requirements for successful pre-training. Surprisingly, we find that a substantial reduction of synthetic images from 1k to 1 can even lead to an increase in pre-training performance, a motivation to further investigate ''scaling backwards''. Finally, we extend our method from synthetic images to real images to see if a single real image can show similar pre-training effect through shape augmentation. We find that the use of grayscale images and affine transformations allows even real images to ''scale backwards''., Comment: Accepted to ECCV2024

The growing interest in human-grade total body positron emission tomography (PET) systems has also application in small animal research. Due to the existing limitations in human-based studies involving drug development and novel treatment monitoring, animal-based research became a necessary step for testing and protocol preparation. In this simulation-based study two unconventional, cost-effective small animal total body PET scanners (for mouse and rat studies) have been investigated in order to inspect their feasibility for preclinical research. They were designed with the novel technology explored by the Jagiellonian-PET (J-PET) Collaboration. Two main PET characteristics: sensitivity and spatial resolution were mainly inspected to evaluate their performance. Moreover, the impact of the scintillator dimension and time-of-flight on the latter parameter was examined in order to design the most efficient tomographs. The presented results show that for mouse TB J-PET the achievable system sensitivity is equal to 2.35% and volumetric spatial resolution to 9.46 +- 0.54 mm3, while for rat TB J-PET they are equal to 2.6% and 14.11 +- 0.80 mm3, respectively. Furthermore, it was shown that the designed tomographs are almost parallax-free systems, hence, they resolve the problem of the acceptance criterion tradeoff between enhancing spatial resolution and reducing sensitivity.

The optimisation of the sensitive region of CMOS sensors with complex non-uniform electric fields requires precise simulations, and this can be achieved by a combination of electrostatic field simulations and Monte Carlo methods. This paper presents the guiding principles of such simulations, using a CMOS pixel sensor with a small collection electrode and a high-resistivity epitaxial layer as an example. The full simulation workflow is described, along with possible pitfalls and how to avoid them. For commercial CMOS processes, detailed doping profiles are confidential, but the presented method provides an optimisation tool that is sufficiently accurate to investigate sensor behaviour and trade-offs of different sensor designs without knowledge of proprietary information. The workflow starts with detailed electric field finite element method simulations in TCAD, using generic doping profiles. Examples of the effect of varying different parameters of the simulated sensor are shown, as well as the creation of weighting fields, and transient pulse simulations. The fields resulting from TCAD simulations can be imported into the Allpix Squared Monte Carlo simulation framework, which enables high-statistics simulations, including modelling of stochastic fluctuations from the underlying physics processes of particle interaction. Example Monte Carlo simulation setups are presented and the different parts of a simulation chain are described. Simulation studies from small collection electrode CMOS sensors are presented, and example results are shown for both single sensors and multiple sensors in a test beam telescope configuration. The studies shown are those typically performed on sensor prototypes in test beam campaigns, and a comparison is made to test beam data, showing a maximum deviation of 4% and demonstrating that the approach is viable for generating realistic results., Comment: 22 pages, 25 figures, submitted to Nuclear Instruments and Methods in Physics Research, Section A

Plastic scintillator strips are considered as one of the promising solutions for the cost-effective construction of total-body positron emission tomography, (PET) system. The purpose of the performed measurements is to compare the transparency of long plastic scintillators with dimensions 6 mm x 24 mm x 1000 mm and with all surfaces polished. Six different types of commercial, general purpose, blue-emitting plastic scintillators with low attenuation of visible light were tested, namely: polyvinyl toluene-based BC-408, EJ-200, RP-408, and polystyrene-based Epic, SP32 and UPS-923A. For determination of the best type of plastic scintillator for totalbody Jagiellonian positron emission tomograph (TB-J-PET) construction, emission and transmission spectra, and technical attenuation length (TAL) of blue light-emitting by the scintillators were measured and compared. The TAL values were determined with the use of UV lamp as excitation source, and photodiode as light detector. Emission spectra of investigated scintillators have maxima in the range from 420 nm to 429 nm. The BC-408 and EJ-200 have the highest transmittance values of about 90% at the maximum emission wavelength measured through a 6 mm thick scintillator strip and the highest technical attenuation length reaching about 2000 mm, allowing assembly of long detection modules for time-of-flight (TOF) J-PET scanners. Influence of the 6 mm x 6 mm, 12 mm x 6 mm, 24 mm x 6 mm cross-sections of the 1000 mm long EJ-200 plastic scintillator on the TAL and signal intensity was measured. The highest TAL value was determined for samples with 24 mm x 6 mm cross-section., Comment: 10 pages, 5 figures

The weighted region problem is the problem of finding the weighted shortest path on a plane consisting of polygonal regions with different weights. For the case when the plane is tessellated by squares, we can solve the problem approximately by finding the shortest path on a grid graph defined by placing a vertex at the center of each grid. In this note, we show that the obtained path admits $(\sqrt{2}+1)$-approximation. This improves the previous result of $2\sqrt{2}$.

Humans and many animals show remarkably adaptive behavior and can respond differently to the same input depending on their internal goals. The brain not only represents the intermediate abstractions needed to perform a computation but also actively maintains a representation of the computation itself (task abstraction). Such separation of the computation and its abstraction is associated with faster learning, flexible decision-making, and broad generalization capacity. We investigate if such benefits might extend to neural networks trained with task abstractions. For such benefits to emerge, one needs a task inference mechanism that possesses two crucial abilities: First, the ability to infer abstract task representations when no longer explicitly provided (task inference), and second, manipulate task representations to adapt to novel problems (task recomposition). To tackle this, we cast task inference as an optimization problem from a variational inference perspective and ground our approach in an expectation-maximization framework. We show that gradients backpropagated through a neural network to a task representation layer are an efficient heuristic to infer current task demands, a process we refer to as gradient-based inference (GBI). Further iterative optimization of the task representation layer allows for recomposing abstractions to adapt to novel situations. Using a toy example, a novel image classifier, and a language model, we demonstrate that GBI provides higher learning efficiency and generalization to novel tasks and limits forgetting. Moreover, we show that GBI has unique advantages such as preserving information for uncertainty estimation and detecting out-of-distribution samples.

In the state-of-the-art Positron Emission Tomography (PET), information about the polarization of annihilation photons is not available. Current PET systems track molecules labeled with positron-emitting radioisotopes by detecting the propagation direction of two photons from positron-electron annihilation. However, annihilation photons carry more information than just the site where they originated. Here we present a novel J-PET scanner built from plastic scintillators, in which annihilation photons interact predominantly via the Compton effect, providing information about photon polarization in addition to information on photon direction of propagation. Theoretically, photons from the decay of positronium in a vacuum are maximally entangled in polarization. However, in matter, when the positron from positronium annihilates with the electron bound to the atom, the question arises whether the photons from such annihilation are maximally entangled. In this work, we determine the distribution of the relative angle between polarization orientations of two photons from positron-electron annihilation in a porous polymer. Contrary to prior results for positron annihilation in aluminum and copper, where the strength of observed correlations is as expected for maximally entangled photons, our results show a significant deviation. We demonstrate that in porous polymer, photon polarization correlation is weaker than for maximally entangled photons but stronger than for separable photons. The data indicate that more than 40% of annihilations in Amberlite resin lead to a non-maximally entangled state. Our result indicates the degree of correlation depends on the annihilation mechanism and the molecular arrangement. We anticipate that the introduced Compton interaction-based PET system opens a promising perspective for exploring polarization correlations in PET as a novel diagnostic indicator.

This paper introduces LLM-jp, a cross-organizational project for the research and development of Japanese large language models (LLMs). LLM-jp aims to develop open-source and strong Japanese LLMs, and as of this writing, more than 1,500 participants from academia and industry are working together for this purpose. This paper presents the background of the establishment of LLM-jp, summaries of its activities, and technical reports on the LLMs developed by LLM-jp. For the latest activities, visit https://llm-jp.nii.ac.jp/en/.

We present a new hybrid semi-implicit finite volume / finite element numerical scheme for the solution of incompressible and weakly compressible media. From the continuum mechanics model proposed by Godunov, Peshkov and Romenski (GPR), we derive the incompressible GPR formulation as well as a weakly compressible GPR system. As for the original GPR model, the new formulations are able to describe different media, from elastoplastic solids to viscous fluids, depending on the values set for the model's relaxation parameters. Then, we propose a new numerical method for the solution of both models based on the splitting of the original systems into three subsystems: one containing the convective part and non-conservative products, a second subsystem for the source terms of the distortion tensor and heat flux equations and, finally, a pressure subsystem. In the first stage of the algorithm, the transport subsystem is solved by employing an explicit finite volume method, while the source terms are solved implicitly. Next, the pressure subsystem is implicitly discretised using finite elements. Within this methodology, unstructured grids are employed, with the pressure defined in the primal grid and the rest of the variables computed in the dual grid. To evaluate the performance of the proposed scheme, a numerical convergence analysis is carried out, which confirms the second order of accuracy in space. A wide range of benchmarks is reproduced for the incompressible and weakly compressible cases, considering both solid and fluid media. These results demonstrate the good behaviour and robustness of the proposed scheme in a variety of scenarios and conditions.

We explain the algorithm employed in SHADOW4 to generate bending magnet sources. We provide an overview of the fundamental equations used to calculate the spectral and angular distributions of synchrotron emission from bending magnets. We outline the procedures for ray sampling. We present examples for intensity and polarization of the bending magnet at the old ESRF1 storage ring, and phase spaces for the bending magnet of the upcoming ALSU ring. Finally, we discuss the calculation of the effective source size and compare it with ray tracing., Comment: 6 pages, 9 figures

The Takagi-Taupin equations are solved in their simplest form (zero deformation) to obtain the Bragg-diffracted and transmitted complex amplitudes. The case of plane-parallel crystal plates is discussed using a matrix model. The equations are implemented in an open-source python library crystalpy adapted for numerical applications such as crystal reflectivity calculations and ray tracing., Comment: 11 pages, 6 figures

Optical surfaces represented by second-degree polynomials (quadratic or conics) are ubiquitous in optics. We revisit the equations of the conic shapes in the context of grazing incidence optics, gathering together the curves commonly used in x-ray instruments and synchrotron beamlines. We present the equations for paraboloids, ellipsoids, and hyperboloids in a common and consistent notation. We develop the transformations from centered systems that are commonly used to describe conics and their axes of symmetry, to local coordinate systems centered on the off-axis mirror surfaces. The equations presented are directly applicable to ray tracing, fabrication, and metrology calculations. They can also be used to study misalignments, movement tolerances, and aberrations of optical surfaces., Comment: 16 pages, 6 figures, 5 Tables