Meng Xia, Meenal K. Kheterpal, Samantha C. Wong, Christine Park, William Ratliff, Lawrence Carin, Ricardo Henao
We consider machine-learning-based malignancy prediction and lesion identification from clinical dermatological images, which can be indistinctly acquired via smartphone or dermoscopy capture. Additionally, we do not assume that images contain single lesions, thus the framework supports both focal or wide-field images. Specifically, we propose a two-stage approach in which we first identify all lesions present in the image regardless of sub-type or likelihood of malignancy, then it estimates their likelihood of malignancy, and through aggregation, it also generates an image-level likelihood of malignancy that can be used for high-level screening processes. Further, we consider augmenting the proposed approach with clinical covariates (from electronic health records) and publicly available data (the ISIC dataset). Comprehensive experiments validated on an independent test dataset demonstrate that i) the proposed approach outperforms alternative model architectures; ii) the model based on images outperforms a pure clinical model by a large margin, and the combination of images and clinical data does not significantly improves over the image-only model; and iii) the proposed framework offers comparable performance in terms of malignancy classification relative to three board certified dermatologists with different levels of experience.
Meng Xiao, Xueqin Huang, Anan Fang, C. T. Chan
We build an effective medium theory for two-dimensional photonic crystals comprising a rectangular lattice of dielectric cylinders with the incident electric field polarized along the axis of the cylinders. In particular, we discuss the feasibility of constructing an effective medium theory for the case where the Bloch wave vector is far away from the center of Brillouin zone, where the optical response of the photonic crystal is necessarily anisotropic and hence the effective medium description becomes inevitability angle dependent. We employ the scattering theory and treat the two-dimensional system as a stack of one-dimensional arrays. We consider only the zero-order interlayer diffraction and all the higher order diffraction terms of interlayer scattering are ignored. This approximation works well when the higher order diffraction terms are all evanescent waves and the interlayer distance is far enough for them to decay out. Scattering theory enables the calculation of transmission and reflection coefficients of a finite sized slab, and we extract the effective parameters such as the impedance ($Z_e$) and the refractive index ($n_e$) using a parameter retrieval method. We note that $n_e$ is uniquely defined only in a very limited region of the reciprocal space. ($n_e k_0 a<<1$, where $k_0$ is the wave vector inside the vacuum and a is thickness of the slab for retrieval), but $Z_e$ is uniquely defined and has a well-defined meaning inside a much larger domain in the reciprocal space. For a lossless system, the effective impedance $Z_e$ is purely real for the pass band and purely imaginary in the band gaps. Using the sign of the imaginary part of $Z_e$, we can classify the band gaps into two groups and this classification explains why there is usually no surface state on the boundary of typical fully gapped photonic crystals comprising of a lattice of dielectric cylinders.
Meng Xiao, Wen-Jie Chen, Wen-Yu He, Z. Q. Zhang, C. T. Chan
Inspired by the discovery of quantum hall effect and topological insulator, topological properties of classical waves start to draw worldwide attention. Topological non-trivial bands characterized by non-zero Chern numbers are realized with external magnetic field induced time reversal symmetry breaking or dynamic modulation. Due to the absence of Faraday-like effect, the breaking of time reversal symmetry in an acoustic system is commonly realized with moving background fluids, and hence drastically increases the engineering complexity. Here we show that we can realize effective inversion symmetry breaking and effective gauge field in a reduced two-dimensional system by structurally engineering interlayer couplings, achieving an acoustic analog of the topological Haldane model. We then find and demonstrate unidirectional backscattering immune edge states. We show that the synthetic gauge field is closely related to the Weyl points in the three-dimensional band structure.
Xunxin Cai, Meng Xiao, Zhiyuan Ning, Yuanchun Zhou
In addressing the imbalanced issue of data within the realm of Natural Language Processing, text data augmentation methods have emerged as pivotal solutions. This data imbalance is prevalent in the research proposals submitted during the funding application process. Such imbalances, resulting from the varying popularity of disciplines or the emergence of interdisciplinary studies, significantly impede the precision of downstream topic models that deduce the affiliated disciplines of these proposals. At the data level, proposals penned by experts and scientists are inherently complex technological texts, replete with intricate terminologies, which augmenting such specialized text data poses unique challenges. At the system level, this, in turn, compromises the fairness of AI-assisted reviewer assignment systems, which raises a spotlight on solving this issue. This study leverages large language models (Llama V1) as data generators to augment research proposals categorized within intricate disciplinary hierarchies, aiming to rectify data imbalances and enhance the equity of expert assignments. We first sample within the hierarchical structure to find the under-represented class. Then we designed a prompt for keyword-based research proposal generation. Our experiments attests to the efficacy of the generated data, demonstrating that research proposals produced using the prompts can effectively address the aforementioned issues and generate high quality scientific text data, thus help the model overcome the imbalanced issue.
Mengyi Huang, Meng Xiao, Ludi Wang, Yi Du
Continuous Relation Extraction (CRE) aims to incrementally learn relation knowledge from a non-stationary stream of data. Since the introduction of new relational tasks can overshadow previously learned information, catastrophic forgetting becomes a significant challenge in this domain. Current replay-based training paradigms prioritize all data uniformly and train memory samples through multiple rounds, which would result in overfitting old tasks and pronounced bias towards new tasks because of the imbalances of the replay set. To handle the problem, we introduce the DecouPled CRE (DP-CRE) framework that decouples the process of prior information preservation and new knowledge acquisition. This framework examines alterations in the embedding space as new relation classes emerge, distinctly managing the preservation and acquisition of knowledge. Extensive experiments show that DP-CRE significantly outperforms other CRE baselines across two datasets.
Guojiao Lin, Zhen Meng, Dongjie Wang, Qingqing Long, Yuanchun Zhou, Meng Xiao
Multimodal recommendation systems (MMRS) have received considerable attention from the research community due to their ability to jointly utilize information from user behavior and product images and text. Previous research has two main issues. First, many long-tail items in recommendation systems have limited interaction data, making it difficult to learn comprehensive and informative representations. However, past MMRS studies have overlooked this issue. Secondly, users' modality preferences are crucial to their behavior. However, previous research has primarily focused on learning item modality representations, while user modality representations have remained relatively simplistic.To address these challenges, we propose a novel Graphs and User Modalities Enhancement (GUME) for long-tail multimodal recommendation. Specifically, we first enhance the user-item graph using multimodal similarity between items. This improves the connectivity of long-tail items and helps them learn high-quality representations through graph propagation. Then, we construct two types of user modalities: explicit interaction features and extended interest features. By using the user modality enhancement strategy to maximize mutual information between these two features, we improve the generalization ability of user modality representations. Additionally, we design an alignment strategy for modality data to remove noise from both internal and external perspectives. Extensive experiments on four publicly available datasets demonstrate the effectiveness of our approach.
Jia-Zheng Li, Kai Bai, Cheng Guo, Tian-Rui Liu, Liang Fang, Duanduan Wan, Meng Xiao
Degeneracy points in non-Hermitian systems are of great interest. While a homotopic framework exists for understanding their behavior in the absence of symmetry, it does not apply to symmetry-protected degeneracy points with reduced codimension. In this work, utilizing algebraic topology, we provide a systematic classification of these symmetry-protected degenerate points and investigate the braid conservation rule followed by them. Using a model Hamiltonian and circuit simulation, we discover that, contrary to simple annihilation, pairwise-created symmetry-protected degeneracy points merge into a higher-order degeneracy point, which goes beyond the abelian picture. Our findings empower researchers across diverse fields to uncover new phenomena and applications harnessing symmetry-protected non-Hermitian degeneracy points.
Han-Rong Xia, Jia-Zheng Li, Si-Yu Yuan, Meng Xiao
Higher-order topological insulators, as a novel family of topological phases, are a hot frontier in condensed matter physics due to their adherence to unconventional bulk-boundary correspondence. A three-dimensional second-order topological insulator can support one-dimensional modes along its hinges (dubbed as hinge states). Here, we present a simple and direct method to construct chiral hinge modes based on a Chern-insulator stack. We analyze the existence of the hinge modes through the nontrivial quadrupole indices, and then design a photonic crystal to realize the specific flowing pattern of the hinge mode in our model. The experimental results align well with full-wave simulations, clearly demonstrating the existence of chiral hinge states. We also verify the robustness of these hinge states against defects in our photonic system.
Xiaohan Huang, Dongjie Wang, Zhiyuan Ning, Ziyue Qiao, Qingqing Long, Haowei Zhu, Min Wu, Yuanchun Zhou, Meng Xiao
Tabular data optimization methods aim to automatically find an optimal feature transformation process that generates high-value features and improves the performance of downstream machine learning tasks. Current frameworks for automated feature transformation rely on iterative sequence generation tasks, optimizing decision strategies through performance feedback from downstream tasks. However, these approaches fail to effectively utilize historical decision-making experiences and overlook potential relationships among generated features, thus limiting the depth of knowledge extraction. Moreover, the granularity of the decision-making process lacks dynamic backtracking capabilities for individual features, leading to insufficient adaptability when encountering inefficient pathways, adversely affecting overall robustness and exploration efficiency. To address the limitations observed in current automatic feature engineering frameworks, we introduce a novel method that utilizes a feature-state transformation graph to effectively preserve the entire feature transformation journey, where each node represents a specific transformation state. During exploration, three cascading agents iteratively select nodes and idea mathematical operations to generate new transformation states. This strategy leverages the inherent properties of the graph structure, allowing for the preservation and reuse of valuable transformations. It also enables backtracking capabilities through graph pruning techniques, which can rectify inefficient transformation paths. To validate the efficacy and flexibility of our approach, we conducted comprehensive experiments and detailed case studies, demonstrating superior performance in diverse scenarios.
Meng Xiao, Qian Lin, Shanhui Fan
We report the existence of Weyl points in a class of non-central symmetric metamaterials, which has time reversal symmetry, but does not have inversion symmetry due to chiral coupling between electric and magnetic fields. This class of metamaterial exhibits either type-I or type-II Weyl points depending on its non-local response. We also provide a physical realization of such metamaterial consisting of an array of metal wires in the shape of elliptical helixes which exhibits type-II Weyl points.
Meng Xia, Reshika Palaniyappan Velumani, Yong Wang, Huamin Qu, Xiaojuan Ma
With the rapid development of online education in recent years, there has been an increasing number of learning platforms that provide students with multi-step questions to cultivate their problem-solving skills. To guarantee the high quality of such learning materials, question designers need to inspect how students' problem-solving processes unfold step by step to infer whether students' problem-solving logic matches their design intent. They also need to compare the behaviors of different groups (e.g., students from different grades) to distribute questions to students with the right level of knowledge. The availability of fine-grained interaction data, such as mouse movement trajectories from the online platforms, provides the opportunity to analyze problem-solving behaviors. However, it is still challenging to interpret, summarize, and compare the high dimensional problem-solving sequence data. In this paper, we present a visual analytics system, QLens, to help question designers inspect detailed problem-solving trajectories, compare different student groups, distill insights for design improvements. In particular, QLens models problem-solving behavior as a hybrid state transition graph and visualizes it through a novel glyph-embedded Sankey diagram, which reflects students' problem-solving logic, engagement, and encountered difficulties. We conduct three case studies and three expert interviews to demonstrate the usefulness of QLens on real-world datasets that consist of thousands of problem-solving traces.
Zhaoxi Li, Chunlong Fei, Shenghui Yang, Chenxue Hou, Jianxin Zhao, Yi Li, Chenxi Zheng, Heping Wu, Yi Quan, Tianlong Zhao, Dongdong Chen, Di Li, Gang Niu, Wei Ren, Meng Xiao, Yintang Yang
The manipulation of acoustic waves plays an important role in a wide range of applications. Currently, acoustic wave manipulation typically relies on either acoustic metasurfaces or phased array transducers. The elements of metasurfaces are designed and optimized for a target frequency, which thus limits their bandwidth. Phased array transducers, suffering from high-cost and complex control circuits, are usually limited by the array size and the filling ratio of the control units. In this work, we introduce active coding piezoelectric metasurfaces; demonstrate commonly implemented acoustic wave manipulation functionalities such as beam steering, beam focusing and vortex beam focusing, acoustic tweezers; and eventually realize ultrasound imaging. The information coded on the piezoelectric metasurfaces herein is frequency independent and originates from the polarization directions, pointing either up or down, of the piezoelectric materials. Such a piezoelectric metasurface is driven by a single electrode and acts as a controllable active sound source, which combines the advantages of acoustic metasurfaces and phased array transducers while keeping the devices structurally simple and compact. Our coding piezoelectric metasurfaces can lead to potential technological innovations in underwater acoustic wave modulation, acoustic tweezers, biomedical imaging, industrial non-destructive testing and neural regulation.
Meng Xiao, Dongjie Wang, Min Wu, Pengfei Wang, Yuanchun Zhou, Yanjie Fu
The goal of Feature Selection - comprising filter, wrapper, and embedded approaches - is to find the optimal feature subset for designated downstream tasks. Nevertheless, current feature selection methods are limited by: 1) the selection criteria of these methods are varied for different domains, making them hard to generalize; 2) the selection performance of these approaches drops significantly when processing high-dimensional feature space coupled with small sample size. In light of these challenges, we pose the question: can selected feature subsets be more robust, accurate, and input dimensionality agnostic? In this paper, we reformulate the feature selection problem as a deep differentiable optimization task and propose a new research perspective: conceptualizing discrete feature subsetting as continuous embedding space optimization. We introduce a novel and principled framework that encompasses a sequential encoder, an accuracy evaluator, a sequential decoder, and a gradient ascent optimizer. This comprehensive framework includes four important steps: preparation of features-accuracy training data, deep feature subset embedding, gradient-optimized search, and feature subset reconstruction. Specifically, we utilize reinforcement feature selection learning to generate diverse and high-quality training data and enhance generalization. By optimizing reconstruction and accuracy losses, we embed feature selection knowledge into a continuous space using an encoder-evaluator-decoder model structure. We employ a gradient ascent search algorithm to find better embeddings in the learned embedding space. Furthermore, we reconstruct feature selection solutions using these embeddings and select the feature subset with the highest performance for downstream tasks as the optimal subset.
Tao Liu, Kai Bai, Yicheng Zhang, Duanduan Wan, Yun Lai, C. T. Chan, Meng Xiao
Boundary modes localized on the boundaries of a finite-size lattice experience a finite size effect (FSE) that could result in unwanted couplings, crosstalks and formation of gaps even in topological boundary modes. It is commonly believed that the FSE decays exponentially with the size of the system and thus requires many lattices before eventually becoming negligibly small. Here we identify a special type of FSE of some boundary modes that apparently vanishes at some particular wave vectors along the boundary. Meanwhile, the number of wave vectors where the FSE vanishes equals the number of lattices across the strip. We analytically prove this type of FSE in a simple model and prove this peculiar feature. We also provide a physical system consisting of a plasmonic sphere array where this FSE is present. Our work points to the possibility of almost arbitrarily tunning of the FSE, which facilitates unprecedented manipulation of the coupling strength between modes or channels such as the integration of multiple waveguides and photonic non-abelian braiding.
Meng Xiao, Ziyue Qiao, Yanjie Fu, Hao Dong, Yi Du, Pengyang Wang, Hui Xiong, Yuanchun Zhou
The peer merit review of research proposals has been the major mechanism for deciding grant awards. However, research proposals have become increasingly interdisciplinary. It has been a longstanding challenge to assign interdisciplinary proposals to appropriate reviewers, so proposals are fairly evaluated. One of the critical steps in reviewer assignment is to generate accurate interdisciplinary topic labels for proposal-reviewer matching. Existing systems mainly collect topic labels manually generated by principal investigators. However, such human-reported labels can be non-accurate, incomplete, labor intensive, and time costly. What role can AI play in developing a fair and precise proposal reviewer assignment system? In this study, we collaborate with the National Science Foundation of China to address the task of automated interdisciplinary topic path detection. For this purpose, we develop a deep Hierarchical Interdisciplinary Research Proposal Classification Network (HIRPCN). Specifically, we first propose a hierarchical transformer to extract the textual semantic information of proposals. We then design an interdisciplinary graph and leverage GNNs for learning representations of each discipline in order to extract interdisciplinary knowledge. After extracting the semantic and interdisciplinary knowledge, we design a level-wise prediction component to fuse the two types of knowledge representations and detect interdisciplinary topic paths for each proposal. We conduct extensive experiments and expert evaluations on three real-world datasets to demonstrate the effectiveness of our proposed model.
Meng Xia, Yankun Zhao, Mehmet Hamza Erol, Jihyeong Hong, Juho Kim
With the rise of the gig economy, online language tutoring platforms are becoming increasingly popular. They provide temporary and flexible jobs for native speakers as tutors and allow language learners to have one-on-one speaking practices on demand. However, the lack of stable relationships hinders tutors and learners from building long-term trust. "Distributed tutorship" -- temporally discontinuous learning experience with different tutors -- has been underexplored yet has many implications for modern learning platforms. In this paper, we analyzed tutorship sequences of 15,959 learners and found that around 40% of learners change to new tutors every session; 44% learners change to new tutors while reverting to previous tutors sometimes; only 16% learners change to new tutors and then fix on one tutor. We also found suggestive evidence that higher distributedness -- higher diversity and lower continuity in tutorship -- is correlated to slower improvements in speaking performance scores with a similar number of sessions. We further surveyed 519 and interviewed 40 learners and found that more learners preferred fixed tutorship while some do not have it due to various reasons. Finally, we conducted semi-structured interviews with three tutors and one product manager to discuss the implications for improving the continuity in learning under distributed tutorship.
Meng Xia, Yankun Zhao, Jihyeong Hong, Mehmet Hamza Erol, Taewook Kim, Juho Kim
With the rise of the gig economy, online language tutoring platforms are becoming increasingly popular. These platforms provide temporary and flexible jobs for native speakers as tutors and allow language learners to have one-on-one speaking practices on demand, on which learners occasionally practice the language with different tutors. With such distributed tutorship, learners can hold flexible schedules and receive diverse feedback. However, learners face challenges in consistently tracking their learning progress because different tutors provide feedback from diverse standards and perspectives, and hardly refer to learners' previous experiences with other tutors. We present RLens, a visualization system for facilitating learners' learning progress reflection by grouping different tutors' feedback, tracking how each feedback type has been addressed across learning sessions, and visualizing the learning progress. We validate our design through a between-subjects study with 40 real-world learners. Results show that learners can successfully analyze their progress and common language issues under distributed tutorship with RLens, while most learners using the baseline interface had difficulty achieving reflection tasks. We further discuss design considerations of computer-aided systems for supporting learning under distributed tutorship.
Weiliang Zhang, Zhen Meng, Dongjie Wang, Min Wu, Kunpeng Liu, Yuanchun Zhou, Meng Xiao
Recent advancements in single-cell genomics necessitate precision in gene panel selection to interpret complex biological data effectively. Those methods aim to streamline the analysis of scRNA-seq data by focusing on the most informative genes that contribute significantly to the specific analysis task. Traditional selection methods, which often rely on expert domain knowledge, embedded machine learning models, or heuristic-based iterative optimization, are prone to biases and inefficiencies that may obscure critical genomic signals. Recognizing the limitations of traditional methods, we aim to transcend these constraints with a refined strategy. In this study, we introduce an iterative gene panel selection strategy that is applicable to clustering tasks in single-cell genomics. Our method uniquely integrates results from other gene selection algorithms, providing valuable preliminary boundaries or prior knowledge as initial guides in the search space to enhance the efficiency of our framework. Furthermore, we incorporate the stochastic nature of the exploration process in reinforcement learning (RL) and its capability for continuous optimization through reward-based feedback. This combination mitigates the biases inherent in the initial boundaries and harnesses RL's adaptability to refine and target gene panel selection dynamically. To illustrate the effectiveness of our method, we conducted detailed comparative experiments, case studies, and visualization analysis.
Liang Fang, Kai Bai, Cheng Guo, Tian-Rui Liu, Jia-Zheng Li, Meng Xiao
Non-Hermitian systems and their topological singularities, such as exceptional points (EPs), lines, and surfaces, have recently attracted intense interest. The investigation of these exceptional constituents has led to fruitful applications. The responsivity of the eigenvalue diverges at EPs, and chiral state transfer occurs when encircling an EP. Traditionally, it was believed that these exceptional features were unique to non-Hermitian systems requiring gain, loss, or nonreciprocal hopping. Here, we show that these exceptional features are also present in nonlinear Hermitian systems. We consider two coupled resonators with Kerr nonlinearity in one resonator, and no non-Hermitian terms. We identify EP-like points (ELPs) on the eigenspectra where the critical behaviors are the same as those of typical EPs. Additionally, this nonlinear Hermitian system can be mapped to linear non-Hermitian systems, with ELPs corresponding to EPs. We also demonstrate that encirclement around an ELP in the parameter space leads to unique chiral state transfer behavior.
Cong Li, Qingqing Long, Yuanchun Zhou, Meng Xiao
Dec 24, 2024·q-bio.GN·PDF Large language models (LLMs) have demonstrated remarkable advancements, primarily due to their capabilities in modeling the hidden relationships within text sequences. This innovation presents a unique opportunity in the field of life sciences, where vast collections of single-cell omics data from multiple species provide a foundation for training foundational models. However, the challenge lies in the disparity of data scales across different species, hindering the development of a comprehensive model for interpreting genetic data across diverse organisms. In this study, we propose an innovative hybrid approach that integrates the general knowledge capabilities of LLMs with domain-specific representation models for single-cell omics data interpretation. We begin by focusing on genes as the fundamental unit of representation. Gene representations are initialized using functional descriptions, leveraging the strengths of mature language models such as LLaMA-2. By inputting single-cell gene-level expression data with prompts, we effectively model cellular representations based on the differential expression levels of genes across various species and cell types. In the experiments, we constructed developmental cells from humans and mice, specifically targeting cells that are challenging to annotate. We evaluated our methodology through basic tasks such as cell annotation and visualization analysis. The results demonstrate the efficacy of our approach compared to other methods using LLMs, highlighting significant improvements in accuracy and interoperability. Our hybrid approach enhances the representation of single-cell data and offers a robust framework for future research in cross-species genetic analysis.