Searcharxiv⌕ Search

arXiv subjects

Ying Hu

Publications and source records attributed to Ying Hu.

At least 73 records · Page 4Linked to original sources
arXiv2024

TinyLLaVA Factory: A Modularized Codebase for Small-scale Large Multimodal Models

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.

cs.LG↗
arXiv2024

fMRI Exploration of Visual Quality Assessment

Despite significant strides in visual quality assessment, the neural mechanisms underlying visual quality perception remain insufficiently explored. This study employed fMRI to examine brain activity during image quality assessment and identify differences in human processing of images with varying quality. Fourteen healthy participants underwent tasks assessing both image quality and content classification while undergoing functional MRI scans. The collected behavioral data was statistically analyzed, and univariate and functional connectivity analyses were conducted on the imaging data. The findings revealed that quality assessment is a more complex task than content classification, involving enhanced activation in high-level cognitive brain regions for fine-grained visual analysis. Moreover, the research showed the brain's adaptability to different visual inputs, adopting different strategies depending on the input's quality. In response to high-quality images, the brain primarily uses specialized visual areas for precise analysis, whereas with low-quality images, it recruits additional resources including higher-order visual cortices and related cognitive and attentional networks to decode and recognize complex, ambiguous signals effectively. This study pioneers the intersection of neuroscience and image quality research, providing empirical evidence through fMRI linking image quality to neural processing. It contributes novel insights into the human visual system's response to diverse image qualities, thereby paving the way for advancements in objective image quality assessment algorithms.

cs.MM↗
arXiv2024

Personalized Forgetting Mechanism with Concept-Driven Knowledge Tracing

Knowledge Tracing (KT) aims to trace changes in students' knowledge states throughout their entire learning process by analyzing their historical learning data and predicting their future learning performance. Existing forgetting curve theory based knowledge tracing models only consider the general forgetting caused by time intervals, ignoring the individualization of students and the causal relationship of the forgetting process. To address these problems, we propose a Concept-driven Personalized Forgetting knowledge tracing model (CPF) which integrates hierarchical relationships between knowledge concepts and incorporates students' personalized cognitive abilities. First, we integrate the students' personalized capabilities into both the learning and forgetting processes to explicitly distinguish students' individual learning gains and forgetting rates according to their cognitive abilities. Second, we take into account the hierarchical relationships between knowledge points and design a precursor-successor knowledge concept matrix to simulate the causal relationship in the forgetting process, while also integrating the potential impact of forgetting prior knowledge points on subsequent ones. The proposed personalized forgetting mechanism can not only be applied to the learning of specifc knowledge concepts but also the life-long learning process. Extensive experimental results on three public datasets show that our CPF outperforms current forgetting curve theory based methods in predicting student performance, demonstrating CPF can better simulate changes in students' knowledge status through the personalized forgetting mechanism.

cs.AI↗
arXiv2024

Cyclic branched covers of Seifert links and properties related to the $ADE$ link conjecture

In this article we show that all cyclic branched covers of a Seifert link have left-orderable fundamental groups, and therefore admit co-oriented taut foliations and are not $L$-spaces, if and only if it is not an $ADE$ link up to orientation. This leads to a proof of the $ADE$ link conjecture for Seifert links. When $L$ is an $ADE$ link up to orientation, we determine which of its canonical $n$-fold cyclic branched covers $Σ_n(L)$ have non-left-orderable fundamental groups. In addition, we give a topological proof of Ishikawa's classification of strongly quasipositive Seifert links and we determine the Seifert links that are definite, resp. have genus zero, resp. have genus equal to its smooth $4$-ball genus, among others. In the last section, we provide a comprehensive survey of the current knowledge and results concerning the $ADE$ link conjecture.

math.GT↗
arXiv2024

TinyLLaVA: A Framework of Small-scale Large Multimodal Models

We present the TinyLLaVA framework that provides a unified perspective in designing and analyzing the small-scale Large Multimodal Models (LMMs). We empirically study the effects of different vision encoders, connection modules, language models, training data and training recipes. Our extensive experiments showed that better quality of data combined with better training recipes, smaller LMMs can consistently achieve on-par performances compared to bigger LMMs. Under our framework, we train a family of small-scale LMMs. Our best model, TinyLLaVA-3.1B, achieves better overall performance against existing 7B models such as LLaVA-1.5 and Qwen-VL. We hope our findings can serve as baselines for future research in terms of data scaling, training setups and model selections. Our model weights and codes will be made public.

cs.LG↗
arXiv2024

Mean-field backward stochastic differential equations and nonlocal PDEs with quadratic growth

In this paper, we study general mean-field backward stochastic differential equations (BSDEs, for short) with quadratic growth. First, the existence and uniqueness of local and global solutions are proved with some new ideas for a one-dimensional mean-field BSDE when the generator $g\big(t, Y, Z, \mathbb{P}_{Y}, \mathbb{P}_{Z}\big)$ has a quadratic growth in $Z$ and the terminal value is bounded. Second, a comparison theorem for the general mean-field BSDEs is obtained with the Girsanov transform. Third, we prove the convergence of the particle systems to the mean-field BSDEs with quadratic growth, and the convergence rate is also given. Finally, in this framework, we use the mean-field BSDE to provide a probabilistic representation for the viscosity solution of a nonlocal partial differential equation (PDE, for short) as an extended nonlinear Feynman-Kac formula, which yields the existence and uniqueness of the solution to the PDE.

math.PR↗
arXiv2023

Intelligent Robotic Sonographer: Mutual Information-based Disentangled Reward Learning from Few Demonstrations

Ultrasound (US) imaging is widely used for biometric measurement and diagnosis of internal organs due to the advantages of being real-time and radiation-free. However, due to inter-operator variations, resulting images highly depend on the experience of sonographers. This work proposes an intelligent robotic sonographer to autonomously "explore" target anatomies and navigate a US probe to a relevant 2D plane by learning from the expert. The underlying high-level physiological knowledge from experts is inferred by a neural reward function, using a ranked pairwise image comparisons approach in a self-supervised fashion. This process can be referred to as understanding the "language of sonography". Considering the generalization capability to overcome inter-patient variations, mutual information is estimated by a network to explicitly disentangle the task-related and domain features in latent space. The robotic localization is carried out in coarse-to-fine mode based on the predicted reward associated with B-mode images. To validate the effectiveness of the proposed reward inference network, representative experiments were performed on vascular phantoms ("line" target), two types of ex-vivo animal organs (chicken heart and lamb kidney) phantoms ("point" target) and in-vivo human carotids, respectively. To further validate the performance of the autonomous acquisition framework, physical robotic acquisitions were performed on three phantoms (vascular, chicken heart, and lamb kidney). The results demonstrated that the proposed advanced framework can robustly work on a variety of seen and unseen phantoms as well as in-vivo human carotid data.

cs.RO↗
arXiv2023

Comparison theorems for multi-dimensional BSDEs with jumps and applications to constrained stochastic linear-quadratic control

In this paper, we, for the first time, establish two comparison theorems for multi-dimensional backward stochastic differential equations with jumps. Our approach is novel and completely different from the existing results for one-dimensional case. Using these and other delicate tools, we then construct solutions to coupled two-dimensional stochastic Riccati equation with jumps in both standard and singular cases. In the end, these results are applied to solve a cone-constrained stochastic linear-quadratic and a mean-variance portfolio selection problem with jumps. Different from no jump problems, the optimal (relative) state processes may change their signs, which is of course due to the presence of jumps.

math.PR↗
arXiv2023

Constrained monotone mean-variance problem with random coefficients

This paper studies the monotone mean-variance (MMV) problem and the classical mean-variance (MV) problem with convex cone trading constraints in a market with random coefficients. We provide semiclosed optimal strategies and optimal values for both problems via certain backward stochastic differential equations (BSDEs). After noting the links between these BSDEs, we find that the two problems share the same optimal portfolio and optimal value. This generalizes the result of Shen and Zou $[$ SIAM J. Financial Math., 13 (2022), pp. SC99-SC112$]$ from deterministic coefficients to random ones.

q-fin.MF↗
arXiv2023

Fractional Quantum Zeno Effect Emerging from Non-Hermitian Physics

Exploring non-Hermitian phenomenology is an exciting frontier of modern physics. However, the demonstration of a non-Hermitian phenomenon that is quantum in nature has remained elusive. Here, we predict quantum non-Hermitian phenomena: the fractional quantum Zeno (FQZ) effect and FQZ-induced photon antibunching. We consider a quantum optics platform with reservoir engineering, where nonlinear emitters are coupled to a bath of decaying bosonic modes whose own decay rates form band structures. By engineering the dissipation band, the spontaneous emission of emitters can be suppressed by strong dissipation through an algebraic scaling with fractional exponents - the FQZ effect. This fractional scaling originates uniquely from the divergent dissipative density of states near the dissipation band edge, different from the traditional closed-bath context. We find FQZ-induced strong photon antibunching in the steady state of a driven emitter even for weak nonlinearities. Remarkably, we identify that the sub-Poissonian quantum statistics of photons, which has no classical analogs, stems here from the key role of non-Hermiticity. Our setup is experimentally feasible with the techniques used to design lattice models with dissipative couplings.

quant-ph↗
arXiv2023

Non-homogeneous stochastic LQ control with regime switching and random coefficients

This paper is concerned with a general non-homogeneous stochastic linear quadratic (LQ) control problem with regime switching and random coefficients. We obtain the explicit optimal state feedback control and optimal value for this problem in terms of two systems of backward stochastic differential equations (BSDEs): one is the famous stochastic Riccati equation and the other one is a new linear multi-dimensional BSDE with all coefficients being unbounded. The existence and uniqueness of the solutions to these two systems of BSDEs are proved by means of BMO martingales and contraction mapping method. At last, the theory is applied to study an asset-liability management problem under the mean-variance criteria.

math.OC↗
arXiv2023

Topological atomic spinwave lattices by dissipative couplings

Recent experimental advance in creating dissipative couplings provides a new route for engineering exotic lattice systems and exploring topological dissipation. Using the spatial lattice of atomic spinwaves in a vacuum vapor cell, where purely dissipative couplings arise from diffusion of atoms, we experimentally realize a dissipative version of the Su-Schrieffer-Heeger (SSH) model. We construct the dissipation spectrum of the topological or trivial lattices via electromagnetically-induced-transparency (EIT) spectroscopy. The topological dissipation spectrum is found to exhibit edge modes within a dissipative gap. We validate chiral symmetry of the dissipative SSH couplings, and also probe topological features of the generalized dissipative SSH model. This work paves the way for realizing non-Hermitian topological quantum optics via dissipative couplings.

quant-ph↗
arXiv2022

Backward stochastic differential equations with conditional reflection and related recursive optimal control problems

We introduce a new type of reflected backward stochastic differential equations (BSDEs) for which the reflection constraint is imposed on its main solution component, denoted as $Y$ by convention, but in terms of its conditional expectation $\mathbb{E}[Y_t|\mathcal{G}_{t}]$ on a general sub-filtration $\{\mathcal{G}_{t}\}.$ We thus term such equation as conditionally reflected BSDE (for short, conditional RBSDE). Conditional RBSDE subsumes classical RBSDE with a pointwise reflection barrier, and the recent developed BSDE with a mean reflection constraint, as its two special and extreme cases: they exactly correspond to $\{\mathcal{G}_{t}\}$ being the full filtration to represent complete information, and the degenerated filtration to deterministic scenario, respectively. For conditional RBSDE, we obtain its existence and uniqueness under mild conditions by combining the Snell envelope method with Skorokhod lemma. We also discuss its connection, in the case of linear driver, to a class of optimal stopping problems in presence of partial information. As a by-product, a new version of comparison theorem is obtained. With the help of this connection, we study weak formulations of a class of optimal control problems with reflected recursive functionals by characterizing the related optimal solution and value. Moreover, in the special case of recursive functionals being RBSDE with pointwise reflections, we study the strong formulations of related stochastic backward recursive control and zero-sum games, both in non-Markovian framework, that are of their own interests and have not been fully explored by existing literatures yet.

math.PR↗
arXiv2022

Optimal consumption-investment with coupled constraints on consumption and investment strategies in a regime switching market with random coefficients

This paper studies finite-time optimal consumption-investment problems with power, logarithmic and exponential utilities, in a regime switching market with random coefficients, subject to coupled constraints on the consumption and investment strategies. We provide explicit optimal consumption-investment strategies and optimal values for the problems in terms of the solutions to some diagonally quadratic backward stochastic differential equation (BSDE) systems and linear BSDE systems with unbound coefficients. Some of these BSDEs are new in the literature and solving them is one of the main theoretical contributions of this paper. We accomplish the latter by applying the truncation, approximation technique to get some a priori uniformly lower and upper bounds for their solutions.

math.PR↗
arXiv2022

General Mean Reflected BSDEs

The present paper is devoted to the study of backward stochastic differential equations with mean reflection formulated by Briand et al. [7]. We investigate the solvability of a generalized mean reflected BSDE, whose driver also depends on the distribution of the solution term $Y$. Using a fixed-point argument, BMO martingale theory and the $θ$-method, we establish the existence and uniqueness result for such BSDEs in several typical situations, including the case where the driver is quadratic with bounded or unbounded terminal condition.

math.PR↗
arXiv2022

Optimal control of SDEs with expected path constraints and related constrained FBSDEs

In this paper, we consider optimal control of stochastic differential equations subject to an expected path constraint. The stochastic maximum principle is given for a general optimal stochastic control in terms of constrained FBSDEs. In particular, the compensated process in our adjoint equation is deterministic, which seems to be new in the literature. For the typical case of linear stochastic systems and quadratic cost functionals (i.e., the so-called LQ optimal stochastic control), a verification theorem is established, and the existence and uniqueness of the constrained reflected FBSDEs are also given.

math.OC↗
arXiv2022

Face Identification Proficiency Test Designed Using Item Response Theory

Measures of face-identification proficiency are essential to ensure accurate and consistent performance by professional forensic face examiners and others who perform face-identification tasks in applied scenarios. Current proficiency tests rely on static sets of stimulus items, and so, cannot be administered validly to the same individual multiple times. To create a proficiency test, a large number of items of "known" difficulty must be assembled. Multiple tests of equal difficulty can be constructed then using subsets of items. We introduce the Triad Identity Matching (TIM) test and evaluate it using Item Response Theory (IRT). Participants view face-image "triads" (N=225) (two images of one identity, one image of a different identity) and select the different identity. In Experiment 1, university students (N=197) showed wide-ranging accuracy on the TIM test, and IRT modeling demonstrated that the TIM items span various difficulty levels. In Experiment 2, we used IRT-based item metrics to partition the test into subsets of specific difficulties. Simulations showed that subsets of the TIM items yielded reliable estimates of subject ability. In Experiments 3a and 3b, we found that the student-derived IRT model reliably evaluated the ability of non-student participants and that ability generalized across different test sessions. In Experiment 3c, we show that TIM test performance correlates with other common face-recognition tests. In summary, the TIM test provides a starting point for developing a framework that is flexible and calibrated to measure proficiency across various ability levels (e.g., professionals or populations with face-processing deficits).

cs.CV↗