arXiv subjects

Jia Yue

Publications and source records attributed to Jia Yue.

12 recordsLinked to original sources

arXiv2025

Enhancing Satellite Object Localization with Dilated Convolutions and Attention-aided Spatial Pooling

Object localization in satellite imagery is particularly challenging due to the high variability of objects, low spatial resolution, and interference from noise and dominant features such as clouds and city lights. In this research, we focus on three satellite datasets: upper atmospheric Gravity Waves (GW), mesospheric Bores (Bore), and Ocean Eddies (OE), each presenting its own unique challenges. These challenges include the variability in the scale and appearance of the main object patterns, where the size, shape, and feature extent of objects of interest can differ significantly. To address these challenges, we introduce YOLO-DCAP, a novel enhanced version of YOLOv5 designed to improve object localization in these complex scenarios. YOLO-DCAP incorporates a Multi-scale Dilated Residual Convolution (MDRC) block to capture multi-scale features at scale with varying dilation rates, and an Attention-aided Spatial Pooling (AaSP) module to focus on the global relevant spatial regions, enhancing feature selection. These structural improvements help to better localize objects in satellite imagery. Experimental results demonstrate that YOLO-DCAP significantly outperforms both the YOLO base model and state-of-the-art approaches, achieving an average improvement of 20.95% in mAP50 and 32.23% in IoU over the base model, and 7.35% and 9.84% respectively over state-of-the-art alternatives, consistently across all three satellite datasets. These consistent gains across all three satellite datasets highlight the robustness and generalizability of the proposed approach. Our code is open sourced at https://github.com/AI-4-atmosphere-remote-sensing/satellite-object-localization.

cs.CV↗

arXiv2025

Stochastic Integration on Stochastic Sets of Interval Type and Applications to Mathematical Finance

In the existing works, stochastic sets $\mathbb{B}$ of interval type, along with $\mathbb{B}$-stochastic processes, were introduced within the framework of stochastic analysis. In this paper, we undertake the construction of $\mathbb{B}$-stochastic integration by exploring three novel types of $\mathbb{B}$-stochastic integrals: Stieltjes integrals of $\mathbb{B}$-predictable processes with respect to $\mathbb{B}$-adapted processes with finite variation, stochastic integrals of $\mathbb{B}$-predictable processes with respect to $\mathbb{B}$-inner local martingales, and stochastic integrals of $\mathbb{B}$-predictable processes with respect to $\mathbb{B}$-inner semimartingales. These $\mathbb{B}$-stochastic integrals are exclusively defined on subsets $\mathbb{B}$, with values outside the scope of $\mathbb{B}$ being deemed irrelevant. Additionally, we present several notable consequences, including the relationship between $\mathbb{B}$-stochastic integrals and existing stochastic integrals, as well as Itô's formula for $\mathbb{B}$-inner semimartingales. In the context of models pertaining to uncertain time-horizons in mathematical finance, we establish essentials of mathematical finance for general markets characterized by sudden-stop horizons. This is achieved by defining self-financing strategies, admissible strategies, and no-arbitrary conditions. In such financial markets, the exclusivity characteristic inherent in $\mathbb{B}$-stochastic integrals offers investors a viable alternative approach. This approach enables them to effectively filter out unnecessary information pertaining to asset price dynamics and portfolio strategies that extend beyond the predefined time-horizons.

math.PR↗

arXiv2024

gWaveNet: Classification of Gravity Waves from Noisy Satellite Data using Custom Kernel Integrated Deep Learning Method

Atmospheric gravity waves occur in the Earths atmosphere caused by an interplay between gravity and buoyancy forces. These waves have profound impacts on various aspects of the atmosphere, including the patterns of precipitation, cloud formation, ozone distribution, aerosols, and pollutant dispersion. Therefore, understanding gravity waves is essential to comprehend and monitor changes in a wide range of atmospheric behaviors. Limited studies have been conducted to identify gravity waves from satellite data using machine learning techniques. Particularly, without applying noise removal techniques, it remains an underexplored area of research. This study presents a novel kernel design aimed at identifying gravity waves within satellite images. The proposed kernel is seamlessly integrated into a deep convolutional neural network, denoted as gWaveNet. Our proposed model exhibits impressive proficiency in detecting images containing gravity waves from noisy satellite data without any feature engineering. The empirical results show our model outperforms related approaches by achieving over 98% training accuracy and over 94% test accuracy which is known to be the best result for gravity waves detection up to the time of this work. We open sourced our code at https://rb.gy/qn68ku.

cs.CV↗

arXiv2024

Modeling of Nitric Oxide Infrared radiative flux in lower thermosphere: a machine learning perspective

Nitric Oxide (NO) significantly impacts energy distribution and chemical processes in the mesosphere and lower thermosphere (MLT). During geomagnetic storms, a substantial influx of energy in the thermosphere leads to an increase in NO infrared emissions. Accurately predicting the radiative flux of Nitric Oxide is crucial for understanding the thermospheric energy budget, particularly during extreme space weather events. With advancements in computational techniques, machine learning (ML) has become a highly effective tool for space weather forecasting. This effort becomes even more worthwhile considering the availability of two decades of continuous NO infrared emissions measurement by TIMED/SABER along with several other key thermospheric variables. We present the scheme of development of an ML-based predictive model for Nitric Oxide Infrared Radiative Flux (NOIRF). Various ML algorithms have been tested for better predictive ability, and an optimized model (NOEMLM) has been developed for the study of NOIRF. This model is able to extract the underlying relationships between the input features and effectively predict the NOIRF. The NOEMLM predictions have very good agreements with SABER observation during quiet time as well as geomagnetic storms. In comparison with the existing TIEGCM model, NOEMLM has very good performance, especially during extreme space weather conditions. The results of this study suggest that utilizing geomagnetic and space weather indices with ML/AI can serve as superior parameters for studying the upper atmosphere, as compared to focusing on specific species having complex chemical processes and associated uncertainties in constituents. ML techniques can effectively carry out the analysis with greater ease than traditional chemical studies.

physics.space-ph↗

arXiv2023

Stochastic Integrals on Predictable Sets of Interval Type with Financial Applications

In this paper, by extending the classic stochastic integrals, we investigate three kinds of more general stochastic integrals: Lebesgue-Stieltjes integrals on predictable sets of interval type (in short: PSITs), stochastic integrals on PSITs of predictable processes with respect to local martingales, and stochastic integrals on PSITs of predictable processes with respect to semimartingales. Such stochastic integrals on PSITs are defined only on restricted stochastic subsets, and their values outside the subsets do not matter. Our study reveals that a stochastic integral on a PSIT can be characterized by a coupled sequence of classic stochastic integrals. Furthermore, the Itô's formula for semimartingales on PSITs is developed for stochastic calculus, and stochastic integrals on PSITs can be applied to more general problems in mathematical finance.

math.PR↗

arXiv2021

Asset Prices with Investor Protection and Survival Analysis of Shareholders in the Cross-Sectional Economy

In this paper, we consider a dynamic asset pricing model in a cross-sectional economy with two firms where a controlling shareholder cannot divert output in one firm with perfect investor protection for minority shareholders and where he can divert a fraction of output in the other firm with imperfect protection. After obtaining the parameters of asset prices by solving the shareholders' consumption-portfolio problems in equilibrium, our model features the effect of investor protection and cross-section in the economy. Furthermore, some survival analysis of the shareholders is presented and sufficient conditions on extinction of the shareholders are given in either firm. Our numerical results are in line with some empirical evidence: (i) poorer investor protection in the cross-sectional economy enables the controlling shareholder to hold less shares of the firm with perfect protection and more shares of the firm with imperfect protection, decreases stock gross returns of both firms, increases stock volatilities of both firms, and decreases interest rates of the economy; (ii) compared with the economy with the single relative firm, for the firm with perfect protection, cross-section enables the controlling shareholder to hold less shares, decreases stock returns, increases stock volatilities slightly and decreases interest rates, while for the firm with imperfect protection, cross-section enables the controlling shareholder to hold more shares, increases stock returns and volatilities and increases interest rates.

math.OC↗

arXiv2020

Asset Prices with Investor Protection and Past Information

In this paper, we consider a dynamic asset pricing model in an approximate fractional economy to address empirical regularities related to both investor protection and past information. Our newly developed model features not only in terms with a controlling shareholder who diverts a fraction of the output, but also good (or bad) memory in his budget dynamics which can be well-calibrated by a pathwise way from the historical data. We find that poorer investor protection leads to higher stock holdings of controlling holders, lower gross stock returns, lower interest rates, and lower modified stock volatilities if the ownership concentration is sufficiently high. More importantly, by establishing an approximation scheme for good (bad) memory of investors on the historical market information, we conclude that good (bad) memory would increase (decrease) aforementioned dynamics and reveal that good (bad) memory strengthens (weakens) investor protection for minority shareholder when the ownership concentration is sufficiently high, while good (bad) memory inversely weakens (strengthens) investor protection for minority shareholder when the ownership concentration is sufficiently low. Our model's implications are consistent with a number of interesting facts documented in the recent literature.

q-fin.PR↗

arXiv2019

Modeling e-Learners' Cognitive and Metacognitive Strategy in Comparative Question Solving

Cognitive and metacognitive strategy had demonstrated a significant role in self-regulated learning (SRL), and an appropriate use of strategies is beneficial to effective learning or question-solving tasks during a human-computer interaction process. This paper proposes a novel method combining Knowledge Map (KM) based data mining technique with Thinking Map (TM) to detect learner's cognitive and metacognitive strategy in the question-solving scenario. In particular, a graph-based mining algorithm is designed to facilitate our proposed method, which can automatically map cognitive strategy to metacognitive strategy with raising abstraction level, and make the cognitive and metacognitive process viewable, which acts like a reverse engineering engine to explain how a learner thinks when solving a question. Additionally, we develop an online learning environment system for participants to learn and record their behaviors. To corroborate the effectiveness of our approach and algorithm, we conduct experiments recruiting 173 postgraduate and undergraduate students, and they were asked to complete a question-solving task, such as "What are similarities and differences between array and pointer?" from "The C Programming Language" course and "What are similarities and differences between packet switching and circuit switching?" from "Computer Network Principle" course. The mined strategies patterns results are encouraging and supported well our proposed method.

cs.CY↗

arXiv2018

Learning Unit State Recognition Based on Multi-channel Data Fusion

Despite recent advances in MOOC, the current e-learning systems have advantages of alleviating barriers by time differences, and geographically spatial separation between teachers and students. However, there has been a 'lack of supervision' problem that e-learner's learning unit state(LUS) can't be supervised automatically. In this paper, we present a fusion framework considering three channel data sources: 1) videos/images from a camera, 2) eye movement information tracked by a low solution eye tracker and 3) mouse movement. Based on these data modalities, we propose a novel approach of multi-channel data fusion to explore the learning unit state recognition. We also propose a method to build a learning state recognition model to avoid manually labeling image data. The experiments were carried on our designed online learning prototype system, and we choose CART, Random Forest and GBDT regression model to predict e-learner's learning state. The results show that multi-channel data fusion model have a better recognition performance in comparison with single channel model. In addition, a best recognition performance can be reached when image, eye movement and mouse movement features are fused.

cs.CV↗

arXiv2018

Volatility swaps valuation under stochastic volatility with jumps and stochastic intensity

In this paper, a pricing formula for volatility swaps is delivered when the underlying asset follows the stochastic volatility model with jumps and stochastic intensity. By using Feynman-Kac theorem, a partial integral differential equation is obtained to derive the joint moment generating function of the previous model. Moreover, discrete and continuous sampled volatility swap pricing formulas are given by employing transform techniques and the relationship between two pricing formulas is discussed. Finally, some numerical simulations are reported to support the results presented in this paper.

q-fin.PR↗

arXiv2018

Variance swaps under Lévy process with stochastic volatility and stochastic interest rate in incomplete markets

This paper focuses on the pricing of the variance swap in an incomplete market where the stochastic interest rate and the price of the stock are respectively driven by Cox-Ingersoll-Ross model and Heston model with simultaneous Lévy jumps. By using the equilibrium framework, we obtain the pricing kernel and the equivalent martingale measure. Moreover, under the forward measure instead of the risk neural measure, we give the closed-form solution for the fair delivery price of the discretely sampled variance swap by employing the joint moment generating function of the underlying processes. Finally, we provide some numerical examples to depict that the values of variance swaps not only depend on the stochastic interest rates but also increase in the presence of jump risks.

q-fin.PR↗

arXiv2017

Fractional Wishart Processes and $\varepsilon$-Fractional Wishart Processes with Applications

In this paper, we introduce two new matrix stochastic processes: fractional Wishart processes and $\varepsilon$-fractional Wishart processes with integer indices which are based on the fractional Brownian motions and then extend $\varepsilon$-fractional Wishart processes to the case with non-integer indices. Both of two kinds of processes include classic Wishart processes when the Hurst index $H$ equals $\frac{1}{2}$ and present serial correlation of stochastic processes. Applying $\varepsilon$-fractional Wishart processes to financial volatility theory, the financial models account for the stochastic volatilities of the assets and for the stochastic correlations not only between the underlying assets' returns but also between their volatilities and for stochastic serial correlation of the relevant assets.

math.OC↗