搜索结果：factory

A neutrino factory is a potential successor to the upcoming generation of neutrino oscillation experiments and a possible precursor to next-generation muon colliders. Such a machine would provide a well-characterized beam of $ν_μ$, $\barν_μ$, $ν_e$, and $\barν_e$ neutrinos with comparable statistics. Here we show the sensitivity of a neutrino factory to new oscillation physics scenarios such as vector neutrino non-standard interactions and CPT violation. We study two different potential setups for a neutrino factory with different assumptions on charge identification in the far detector. We find that 10 years of a neutrino factory combined with 10 years of DUNE can improve over most of the current constraints on these scenarios and even over forecasted constraints by 20 years of DUNE. Additionally, we find that a neutrino factory can break degeneracies between the standard oscillation parameters and neutrino non-standard interaction parameters present at DUNE.

Adding sequence parallelism into LLaMA-Factory, we open-sourced 360-LLaMA-Factory at https://github.com/Qihoo360/360-LLaMA-Factory. 360-LLaMA-Factory has received wide recognition and used in models such as Light-R1 arXiv:2503.10460, TinyR1 arXiv:2503.04872, Kaggle AIMO math models and also in large companies' training frameworks. This technical report delves deeper into the different sequence parallel modes behind 360-LLaMA-Factory and discusses our implementation insights.

Long-form factuality evaluation assesses the ability of models to generate accurate, comprehensive responses to short prompts. Existing benchmarks often lack human verification, leading to potential quality issues. To address this limitation, we introduce FACTORY, a large-scale, human-verified prompt set. Developed using a model-in-the-loop approach and refined by humans, FACTORY includes challenging prompts that are fact-seeking, answerable, and unambiguous. We conduct human evaluations on 6 state-of-the-art language models using FACTORY and existing datasets. Our results show that FACTORY is a challenging benchmark: approximately 40% of the claims made in the responses of SOTA models are not factual, compared to only 10% for other datasets. Our analysis identifies the strengths of FACTORY over prior benchmarks, emphasizing its reliability and the necessity for models to reason across long-tailed facts.

Given a known function $f : [0, 1] \mapsto (0, 1)$ and a random but almost surely finite number of independent, Ber$(x)$-distributed random variables with unknown $x \in [0, 1]$, we construct an unbiased, $[0, 1]$-valued estimator of the probability $f(x) \in (0, 1)$. Our estimator is based on so-called debiasing, or randomly truncating a telescopic series of consistent estimators. Constructing these consistent estimators from the coefficients of a particular Bernoulli factory for $f$ yields provable upper and lower bounds for our unbiased estimator. Our result can be thought of as a novel Bernoulli factory with the appealing property that the required number of Ber$(x)$-distributed random variates is independent of their outcomes, and also as constructive example of the so-called $f$-factory.

We present TinyLLaVA Factory, an open-source modular codebase for small-scale large multimodal models (LMMs) with a focus on simplicity of code implementations, extensibility of new features, and reproducibility of training results. Following the design philosophy of the factory pattern in software engineering, TinyLLaVA Factory modularizes the entire system into interchangeable components, with each component integrating a suite of cutting-edge models and methods, meanwhile leaving room for extensions to more features. In addition to allowing users to customize their own LMMs, TinyLLaVA Factory provides popular training recipes to let users pretrain and finetune their models with less coding effort. Empirical experiments validate the effectiveness of our codebase. The goal of TinyLLaVA Factory is to assist researchers and practitioners in exploring the wide landscape of designing and training small-scale LMMs with affordable computational resources.

Modern automated factories increasingly run manufacturing procedures using a matrix of programmable machines, such as 3D printers, interconnected by a programmable transport system, such as a fleet of tabletop robots. To embed a manufacturing procedure into a smart factory, an operator must: (a) assign each of its processes to a machine and (b) specify how agents should transport parts between machines. The problem of embedding a manufacturing process into a smart factory is termed the Smart Factory Embedding (SFE) problem. State-of-the-art SFE solvers can only scale to factories containing a couple dozen machines. Modern smart factories, however, may contain hundreds of machines. We fill this hole by introducing the first highly scalable solution to the SFE, TS-ACES, the Traffic System based Anytime Cyclic Embedding Solver. We show that TS-ACES is complete and can scale to SFE instances based on real industrial scenarios with more than a hundred machines.

In the era of Industry 4.0, smart factories have emerged as a paradigm shift, redefining manufacturing with the integration of advanced digital technologies. Central to this transformation is the deployment of 5G networks, offering unprecedented levels of connectivity, speed, reliability, and ultra-low latency. Among the revolutionary features of 5G is network slicing, a technology that offers enhanced capabilities through the customization of network resources by allowing multiple logical networks (or slices) to run on top of a shared physical infrastructure. This capability is particularly crucial in the densely packed and highly dynamic environment of smart factories, where diverse applications - from robotic automation to real-time analytics - demand varying network requirements. In this paper, we present a comprehensive overview of the integration of slicing in smart factory networks, emphasizing its critical role in enhancing operational efficiency and supporting the diverse requirements of future manufacturing processes. We elaborate on the recent advances, and technical scenarios, including indoor factory propagation conditions, traffic characteristics, system requirements,

The physics program of the Higgs factory will focus on measurements of the 125 GeV Higgs boson, with the Higgs-strahlung process being the dominant production channel at 250 GeV. However, production of extra light scalars is still not excluded by the existing experimental data, provided their coupling to the gauge bosons is sufficiently suppressed. Fermion couplings of such a scalar could also be very different from the SM predictions leading to non-standard decay paterns. Presented in this contribution are results from the ongoing studies on prospects of direct light scalar observation at future Higgs factory experiments in different decay channels.

Manufacturing industry is heading towards socialization, interconnection, and platformization. Motivated by the infiltration of sharing economy usage in manufacturing, this paper addresses a new factory model -- shared factory -- and provides a theoretical architecture and some actual cases for manufacturing sharing. Concepts related to three kinds of shared factories which deal respectively with sharing production-orders, manufacturing-resources and manufacturing-capabilities, are defined accordingly. These three kinds of shared factory modes can be used for building correspondent sharing manufacturing ecosystems. On the basis of sharing economic analysis, we identify feasible key enabled technologies for configuring and running a shared factory. At the same time, opportunities and challenges of enabling the shared factory are also analyzed in detail. In fact, shared factory, as a new production node, enhances the sharing nature of social manufacturing paradigm, fits the needs of light assets and gives us a new chance to use socialized manufacturing resources. It can be drawn that implementing a shared factory would reach a win-win way through production value-added transformation

An electron-positron collider designed for precision studies of the Higgs boson, a so-called Higgs factory is the highest-priority next collider of the particle physics community. This contribution summarises the key physics goals of such a Higgs factory and reviews the status of the various proposed realisations from mature concepts to very recent ideas. The commonalities and special advantages of circular and linear approaches will be discussed, respectively, before highlighting some recent developments regarding the key technologies, the operation scenarios and sustainability aspects for future colliders.

We discuss the optimization of a neutrino factory for large \sin^2 2 θ_{13}, where we assume minimum effort on the accelerator side. This implies that we use low muon energies for the price of an optimized detection system. We demonstrate that such a neutrino factory performs excellent if combined with the electron neutrino appearance channel. Instead of the platinum channel operated with the muon neutrinos from the muon decays, we propose to use the initial superbeam from the decaying pions and kaons, which might be utilized at little extra effort. Since we assume out-of-phase bunches arriving at the same detector, we do not require electron charge identification. In addition, we can choose the proton energy such that we obtain a synergistic spectrum peaking at lower energies. We find that both the superbeam and the neutrino factory beam should used at the identical baseline to reduce matter density uncertainties, possibly with the same detector. This effectively makes the configuration a single experiment, which we call ``neutrino factory superbeam''. We demonstrate that this experiment outperforms a low-energy neutrino factory or a wide band beam alone beyond a simple addition o

An easier and intuitive interface architecture is necessary for digital twin of plant factory. I suggest an immersive and interactive mixed reality interface for digital twin models of smart farming, for remote work rather than simulation of components. The environment is constructed with UI display and a streaming background scene, which is a real time scene taken from camera device located in the plant factory, processed with deformable neural radiance fields. User can monitor and control the remote plant factory facilities with HMD or 2D display based mixed reality environment. This paper also introduces detailed concept and describes the system architecture to implement suggested mixed reality interface.

In this position paper, we present our approach of utilizing mobile devices (i.e., mobile phones and tablets) for assisting engineers and experts in understanding and maintaining the factory pipelines. For this, we present a platform, called assistME, that is composed of three main components: the assistME Server, the assistME mobile infrastructure, and the co-assistME collaborative environment. In order to get full utilization of the assistME platform, we assume that an initial setup is made in the factory in such a way that it is equipped with different sensors to collect data about specific events in the factory pipeline together with the corresponding locations of these events. The assistME Server works as a central control unit in the platform and collects data from the installed sensors in the factory pipeline. In the case of any unexpected behavior or any critical situation in the factory pipeline, notification and other details are sent to the related group of engineers and experts through the assistME mobile app. Further, the co-assistME collaborative environment, equipped with a large shared screen and multiple mobile devices, helps the engineers and experts to collaborat

Despite the great success of the $Υ(4S)$ $B$ factories at KEK and SLAC and the guaranteed addition of high sensitivity measurements on beauty decays to be performed at the Tevatron and LHC, a strong case can be made for an $e^+e^-$ Super-$B$ factory yielding data samples of order $10^{10}$ $B \bar B$ pairs as a necessity rather than luxury. It has to be justified through its ability to not only establish deviations from the Standard Model, but also diagnose and interpret those in terms of specific features of the New Dynamics. The role to be played by a Super-$B$ factory is thus analogous {\em and even in parallel} to that of a linear collider. The latter's goal is to provide more detailed information on previously discovered New Physics involved in the electroweak phase transition. Likewise a Super-$B$ factory would provide precision probes for analyzing whether such New Dynamics has an impact on heavy flavour dynamics -- a need particularly manifest if the New Physics is housed under the `big tent' of SUSY. The huge statistics of a Super-$B$ factory and the comprehensive body of accurate measurements uniquely possible there would be harnessed in several classes of studies, among

We discuss the impact of near detectors at a neutrino factory both on standard oscillation and non-standard interaction measurements. Our systematics treatment includes cross section errors, flux errors, and background uncertainties, and our near detector fluxes include the geometry of the neutrino source and the detector. Instead of a specific detector concept, we introduce qualitatively different classes of near detectors with different characteristics, such as near detectors catching the whole neutrino flux (near detector limit) versus near detectors observing a spectrum similar to that of the far detector (far detector limit). We include the low energy neutrino factory in the discussion. We illustrate for which measurements near detectors are required, discuss how many are needed, and what the role of the flux monitoring is. For instance, we demonstrate that near detectors are mandatory for the leading atmospheric parameter measurements if the neutrino factory has only one baseline, whereas systematical errors partially cancel if the neutrino factory complex includes the magic baseline. Finally, near detectors with nu_tau detection are shown to be useful for non-standard intera

The long-term prospects for fully exploring three-flavor mixing in the neutrino sector depend upon an ongoing and increased investment in the appropriate accelerator R&D. Two new concepts have been proposed that would revolutionize neutrino experiments, namely the Neutrino Factory and the Beta Beam facility. These new facilities would dramatically improve our ability to test the three-flavor mixing framework, measure \textsl{CP} violation in the lepton sector, and perhaps determine the neutrino mass hierarchy, and, if necessary, probe extremely small values of the mixing angle $θ_{13}$. The stunning sensitivity that could be achieved with a Neutrino Factory is described, together with our present understanding of the corresponding sensitivity that might be achieved with a Beta Beam facility. In the Beta Beam case, additional study is required to better understand the optimum Beta Beam energy, and the achievable sensitivity. Neither a Neutrino Factory nor a Beta Beam facility could be built without significant R&D. An impressive Neutrino Factory R&D effort has been ongoing in the U.S. and elsewhere over the last few years and significant progress has been made towards op

Industry 4.0 revolution concerns the digital transformation of manufacturing and promises to answer the ever-increasing demand of product customisation and manufacturing flexibility while incurring low costs. To perform the required factory reconfiguration, a computationally demanding optimisation process has to be executed to find favourable solutions in a relatively short time. While previous research focused on planning and scheduling of smart factories based on cloud-based optimisation, little attention has been paid to effective approaches to describe the targeted factory, the required products and the production processes. However, these matters are fundamental for the optimisation engine to be correctly and efficiently performed. This paper presents an XML-based factory modelling language to effectively describe the above data for a given factory and commodity order and to provide a convenient interface for altering the input information. Finally, two real-world manufacturing plants are provided to illustrate the feasibility of the proposed description language.

A neutrino factory has unparalleled physics reach for the discovery and measurement of CP violation in the neutrino sector. A far detector for a neutrino factory must have good charge identification with excellent background rejection and a large mass. An elegant solution is to construct a magnetized iron neutrino detector (MIND) along the lines of MINOS, where iron plates provide a toroidal magnetic field and scintillator planes provide 3D space points. In this report, the current status of a simulation of a toroidal MIND for a neutrino factory is discussed in light of the recent measurements of large $θ_{13}$. The response and performance using the 10 GeV neutrino factory configuration are presented. It is shown that this setup has equivalent $δ_{CP}$ reach to a MIND with a dipole field and is sensitive to the discovery of CP violation over 85% of the values of $δ_{CP}$.

This year, 2015, marks the centenary of the publication of Einsteins Theory of General Relativity and it has been named the International Year of Light and light-based technologies by the UN General Assembly. It is thus timely to discuss the possibility of broadening the present CERN research program by including a new component based on a novel concept of the light source which could pave a way towards a multipurpose Gamma Factory. The proposed light source could be realized at CERN by using the infrastructure of the existing accelerators. It could push the intensity limits of the presently operating light-sources by at least 7 orders of magnitude, reaching the flux of the order of 10^17 photons/s, in the particularly interesting gamma-ray energy domain of 1 < Ephoton < 400 MeV. This domain is out of reach for the FEL-based light sources. The energy-tuned, quasi-monochromatic gamma beams, together with the gamma-beam-driven, high intensity secondary beams of polarized positrons, polarized muons, neutrons and radioactive ions would constitute the basic research tools of the proposed Gamma Factory. The Gamma Factory could open new research opportunities in a vast domain of unc