2063 research outputs found
Sort by
P-NP instance decomposition based on the Fourier transform for solving the Linear Ordering Problem
No full textThe Fourier transform over finite groups has proved to be a useful tool for analyzing combinatorial optimization problems. However, few heuristic and meta-heuristic algorithms have been proposed in the literature that utilize the information provided by this technique to guide the search process. In this work, we attempt to address this research gap by considering the case study of the Linear Ordering Problem (LOP).
Based on the Fourier transform, we propose an instance decomposition strategy that divides any LOP instance into the sum of two LOP instances associated with a P and an NP-Hard optimization problem. By linearly aggregating the instances obtained from the decomposition, it is possible to create artificial instances with modified proportions of the P and NP-Hard components. Conducted experiments show that increasing the weight of the P component leads to a less rugged fitness landscape suitable for local search-based optimization. We take advantage of this phenomenon by presenting a new meta-heuristic algorithm called P-Descent Search (PDS). The proposed method, first, optimizes a surrogate instance with a high proportion of the P component, and then, gradually increases the weight of the NP-Hard component until the original instance is reached. The multi-start version of PDS shows a promising and predictable performance that appears to be correlated to specific characteristics of the problem, which could open the door to an automatic tuning of its hyper-parameters
Accuracy and Safety Between Robot-Assisted and Conventional Freehand Fluoroscope-Assisted Placement of Pedicle Screws in Thoracolumbar Spine: Meta-Analysis
No full textBackground and Objectives: Robotic-assisted surgery (RS) has progressively emerged as a promising technology in modern thoracolumbar spinal surgery, offering the potential to enhance accuracy and improve clinical outcomes. To date, the benefits of robot-assisted techniques in thoracolumbar spinal surgery remain controversial. The objective of this study was to assess the efficacy and safety of RS compared to fluoroscopyassisted surgery (FS) in spinal fusion procedures. Materials and Methods: In accordance with the PRISMA guidelines, a systematic review and meta-analysis was conducted, using REVMAN V5.3 software. The review protocol was registered in the Prospective Register of Systematic Reviews (PROSPERO) website with the following registration number: CRD42024567193. Results: Eighteen studies were included in the meta-analysis with a total of 1566 patients examined. The results demonstrated a worse accuracy in FS in cases with major violations of the peduncular cortex (D–E grades, according to Gertzbein’s classification) [(odds ratio (OR) 0.47, 95%-CI 0.28 to 0.80, I2 0%]. In addition, a lower complication rate was shown in the RS group compared to the FS group, specifically regarding the need for surgical revision due to screw mispositioning (OR 0.28-CI 0.17 to 0.48, I2 98%). Conclusions: Advantages of robot-assisted techniques were demonstrated in terms of postoperative complications, revision surgery rates, and the accuracy of screw
placement. While RS represents a valuable and promising technological advancement in thoracolumbar spinal surgery, future studies are needed to further explore its advantages in thoracolumbar spinal surgery and to identify which spinal surgical approach has greater advantages when using the robot
Computational Modelling of Thixotropic Multiphase Fluids
No full textMultiphase systems are ubiquitous in engineering, biology, and materials science, where understanding their complex interactions
and rheological behavior is crucial for advancing applications ranging from emulsion stability to cellular phase separation. This
study presents a numerical methodology for modeling thixotropic multiphase fluids, emphasizing the transient behavior of viscosity
and the intricate interactions between phases. The model incorporates phase-dependent viscosities, interfacial tension effects,
and the dynamics of phase separation, coalescence, and break-up, making it suitable for simulating systems with complex flow
regimes. A key feature of the methodology is its ability to capture thixotropic behavior, where viscosity evolves over time due
to microstructural changes induced by shear history. This approach enables the simulation of aging and recovery processes in
materials such as gels, emulsions, and biological tissues. The model is rigorously validated against benchmark cases, demonstrating
its accuracy in predicting multiphase systems under static and dynamic conditions. Subsequently, the methodology is applied to
investigate systems with varying levels of microstructural evolution, revealing the impact of thixotropic dynamics on overall system
behavior. The results provide new insights into the time-dependent rheology of multiphase fluids and highlight the versatility of the
model for applications in industrial and biological systems involving complex fluid interactions
The Gauge Theoretic Structure of Non-Locality and Contextuality
No full textAbramsky and Brandenburger show that nonlocality and contextuality arise when locally compatible data cannot be glued into a single global assignment. In a global topological data space, interactions among locally defined maps can produce patterns that are consistent within each context yet incompatible globally—the empirical hallmark of quantum contextuality. We make this concrete with an evolving simplicial-bundle framework and apply this framework to several standard empirical models in quantum foundations.PID2023-146683OB-100 funded by MICIU/AEI /10.13039/501100011033 and by ERDF, EU
THE INITIAL-TO-FINAL-STATE INVERSE PROBLEM WITH UNBOUNDED POTENTIALS AND STRICHARTZ ESTIMATES
No full textThe initial-to-final-state inverse problem consists in determining
a quantum Hamiltonian assuming the knowledge of the state of the system
at some fixed time, for every initial state. We formulated this problem to
establish a theoretical framework that would explain the viability of data-
driven prediction in quantum mechanics. In a previous work, we analysed this
inverse problem for Hamiltonians of the form−∆+V with an electric potential
V= V (t, x), and we showed that uniqueness holds whenever the potentials are
bounded and decay super-exponentially at infinity. In this paper, we extend
this result for unbounded potentials. One of the key steps consists in proving a
family of suitable Strichartz estimates—including the corresponding endpoint
of Keel and Tao.
In the context of the inverse Calder´on problem this family of inequalities
corresponds to the Carleman inequality proved by Kenig, Ruiz and Sogge.
Haberman showed that this inequality can be also retrieved as an embedding
of a suitable Bourgain space. The corresponding Bourgain space in our context
do not capture the mixed-norm Lebesgue spaces of Strichartz inequalities.
In this paper, we give a counterexample that justifies this fact, and shows
the limitations of Bourgain spaces to address the initial-to-final-state inverse
problem
Smooth surface finishing for 5-axis flank CNC machining of free-form geometries using custom-shaped tools
No full textKK-2024/00024
PID2023-146640NBI00
MICIU/AEI/10.13039/50110001103
Teacher privileged distillation: How to deal with imperfect teachers?
Full text linkThe paradigm of learning using privileged information leverages privileged features present at training time, but not at prediction, as additional training information. The privileged learning process is addressed through a knowledge distillation perspective: information from a teacher learned with regular and privileged features is transferred to a student composed exclusively of regular features. While most approaches assume perfect knowledge for the teacher, it can commit mistakes. Assuming that, we propose a novel privileged distillation framework with a double contribution. Firstly, a designed function to imitate the teacher when it classifies correctly and to differ in cases of misclassification. Secondly, an adaptation of the cross-entropy loss to appropriately penalize the instances where the student outperforms the teacher. Its effectiveness is empirically demonstrated on datasets with imperfect teachers, significantly enhancing the performance of state-of-the-art frameworks. Furthermore, necessary conditions for successful privileged learning are presented, along with a dataset categorization based on the information provided by the privileged features.PID 2022-137442NB-I00
PRE2021-09927
Joint modeling with beta-binomial distribution for patient-reported outcomes and survival data
Full text linkThis thesis addresses critical methodological gaps in the joint analysis of patient-reported outcomes (PROs) and survival data. PROs, as discrete bounded measures with inherent overdispersion, require specialized statistical treatment that conventional Gaussian-based joint models fail to provide. We develop novel methodological frameworks that properly account for PRO characteristics through beta-binomial distributions, overcoming limitations of existing approaches.
In this work, we propose, explore, and discuss various statistical approaches for joint modeling, from frequentist to bayesian proposals. Our work highlights the advantages of joint models that integrate longitudinal and survival data while emphasizing the importance of choosing appropriate distributions for PRO data.
In particular, in this dissertation, we propose three joint models to analyze both, the longitudinal PRO and survival data:
a) A frequentist two-stage approach, providing initial practical solutions. In this proposal, the central innovation lies in a joint model based on a two-stage methodology that incorporates the beta-binomial distribution for the longitudinal submodel. This methodology avoids computational complexities while ensuring a distributional fit that considers the natural characteristics of PRO (discrete, bounded and overdispersed).
b) A Bayesian one-stage joint model, offering improved estimation. In this proposal, the main objective was to keep the distributional features for PRO data regarding beta-binomial distribution while performing a joint specification approach. The Bayesian formulation of the problem allows us to avoid the computational complexities we found in frequentist approaches. Moreover, we considered the parameters’ posterior estimations to perform dynamic predictions of the survival probabilities, being updated as more longitudinal PRO information is considered.
c) A multivariate Bayesian framework, enabling simultaneous analysis of multiple PRO dimensions with survival outcomes. In this proposal, our primary objective was to address the multidimensional structure of questionnaires within the joint modeling framework. However, when dealing with multivariate approaches, it might be necessary to use regularization techniques to avoid possible multicollinearity. Therefore, we explored common regularization techniques in the literature within the joint modeling framework.
The proposed methods' performance is evaluated using simulation studies, and comparisons with common approaches in the literature are provided. Additionally, we applied these methods to analyze a study carried out with chronic obstructive pulmonary disease (COPD) patients, where longitudinal tendencies for PRO data collected and their relationship with patients’ mortality are of interest
Topics in harmonic analysis related to Rubio de Francia square functions and directional singular integrals
Full text linkThis thesis focuses on the -boundedness of maximal directional singular integral operators and Rubio de Francia square functions.
Firstly, we study maximal directional singular integral operators in defined by a Hörmander--Mihlin multiplier on an -dimensional subspace which act trivially in the perpendicular direction. The choice of subspace depends measurably on the first variables of . Assuming the subspace to be non-degenerate in the sense that it is contained in a subspace of away of a cone around and the function to be frequency supported in a cone away from , we prove bounds for these operators when . If we assume, additionally, that is supported in a single frequency band, we are able to extend the boundedness range to . The non-degeneracy assumption cannot, in general, be removed, even in the band-limited case.
Secondly, we study one-dimensional square functions in the spirit of Rubio de Francia. Let be the Fourier restriction of to an interval . If is an arbitrary collection of pairwise disjoint intervals, the square function of is termed the Rubio de Francia square function . In this thesis we prove a pointwise bound for by a sparse operator involving local -averages. A pointwise bound for the smooth version of by a sparse square function is also proved. These pointwise localization principles lead to quantitative , , and weighted weak-type , , norm inequalities for . In particular, the obtained weak norm bounds are new for and sharp for . The proofs rely on sparse bounds for abstract balayages of Carleson sequences, oscillation inequalities, local orthogonality and time-frequency analysis discretization techniques. The thesis also contains two results related to the outstanding conjecture that is bounded on if and only if . The conjecture is verified for radially decreasing, even weights, and in full generality for the Walsh group analogue of .predoc Basque Government grant 2022 "Programa Predoctoral de Formación de Personal Investigador No Doctor",
PID2021-122156NB-I00,
PID2023-146646NB-I0
Self-Composing Policies for Scalable Continual Reinforcement Learning
Full text linkThis work introduces a growable and modular
neural network architecture that naturally avoids
catastrophic forgetting and interference in con-
tinual reinforcement learning. The structure of
each module allows the selective combination of
previous policies along with its internal policy,
accelerating the learning process on the current
task. Unlike previous growing neural network
approaches, we show that the number of parame-
ters of the proposed approach grows linearly with
respect to the number of tasks, and does not sac-
rifice plasticity to scale. Experiments conducted
in benchmark continuous control and visual prob-
lems reveal that the proposed approach achieves
greater knowledge transfer and performance than
alternative method