arXiv subjects

Qianli Ma

Publications and source records attributed to Qianli Ma.

At least 109 records · Page 6Linked to original sources

arXiv2023

Machine Learning Interpretability of Outer Radiation Belt Enhancement \& Depletion Events

We investigate the response of outer radiation belt electron fluxes to different solar wind and geomagnetic indices using an interpretable machine learning method. We reconstruct the electron flux variation during 19 enhancement and 7 depletion events and demonstrate a feature attribution analysis on the superposed epoch results for the first time. We find that the intensity and duration of the substorm sequence following an initial dropout determine the overall enhancement or depletion of electron fluxes, while the solar wind pressure drives the initial dropout in both types of events. Further statistical results from a dataset with 71 events confirm this and show a significant correlation between the resulting flux levels and the average AL index, indicating that the observed "depletion" event can be more accurately described as a "non-enhancement" event. Our novel SHAP-Enhanced Superposed Epoch Analysis (SHESEA) method can be used as an insight discovery tool in various physical systems.

physics.space-ph↗

arXiv2023

Field-Induced Partial Disorder in a Shastry-Sutherland Lattice

A 2-Q antiferromagnetic order of the ferromagnetic dimers was found below TN = 2.9 K in the Shastry-Sutherland lattice BaNd2ZnS5 by single crystal neutron diffraction. The magnetic order can be understood by the orthogonal arrangement of local Ising Nd spins, identified by polarized neutrons. A field was applied along [1 -1 0] to probe the observed metamagnetic transition in the magnetization measurement. The field decouples two magnetic sublattices corresponding to the propagation vectors q1= (0.5, 0.5, 0) and q2= (-0.5, 0.5, 0), respectively. Each sublattice shows a stripe order with a Neel-type arrangement in each single layer. The stripe order with q1 remains nearly intact up to 6 T, while the other one with q2 is suppressed at a critical field Hc ~1.7 T, indicating a partial disorder. The Hc varies with temperature and is manifested in the H-T phase diagram constructed by measuring the magnetization in BaNd2ZnS5.

cond-mat.str-el↗

arXiv2023

Preserving Commonsense Knowledge from Pre-trained Language Models via Causal Inference

Fine-tuning has been proven to be a simple and effective technique to transfer the learned knowledge of Pre-trained Language Models (PLMs) to downstream tasks. However, vanilla fine-tuning easily overfits the target data and degrades the generalization ability. Most existing studies attribute it to catastrophic forgetting, and they retain the pre-trained knowledge indiscriminately without identifying what knowledge is transferable. Motivated by this, we frame fine-tuning into a causal graph and discover that the crux of catastrophic forgetting lies in the missing causal effects from the pretrained data. Based on the causal view, we propose a unified objective for fine-tuning to retrieve the causality back. Intriguingly, the unified objective can be seen as the sum of the vanilla fine-tuning objective, which learns new knowledge from target data, and the causal objective, which preserves old knowledge from PLMs. Therefore, our method is flexible and can mitigate negative transfer while preserving knowledge. Since endowing models with commonsense is a long-standing challenge, we implement our method on commonsense QA with a proposed heuristic estimation to verify its effectiveness. In the experiments, our method outperforms state-of-the-art fine-tuning methods on all six commonsense QA datasets and can be implemented as a plug-in module to inflate the performance of existing QA models.

cs.CL↗

arXiv2023

Perturbation-based Self-supervised Attention for Attention Bias in Text Classification

In text classification, the traditional attention mechanisms usually focus too much on frequent words, and need extensive labeled data in order to learn. This paper proposes a perturbation-based self-supervised attention approach to guide attention learning without any annotation overhead. Specifically, we add as much noise as possible to all the words in the sentence without changing their semantics and predictions. We hypothesize that words that tolerate more noise are less significant, and we can use this information to refine the attention distribution. Experimental results on three text classification tasks show that our approach can significantly improve the performance of current attention-based models, and is more effective than existing self-supervised methods. We also provide a visualization analysis to verify the effectiveness of our approach.

cs.CL↗

arXiv2023

Dynamic Point Fields

Recent years have witnessed significant progress in the field of neural surface reconstruction. While the extensive focus was put on volumetric and implicit approaches, a number of works have shown that explicit graphics primitives such as point clouds can significantly reduce computational complexity, without sacrificing the reconstructed surface quality. However, less emphasis has been put on modeling dynamic surfaces with point primitives. In this work, we present a dynamic point field model that combines the representational benefits of explicit point-based graphics with implicit deformation networks to allow efficient modeling of non-rigid 3D surfaces. Using explicit surface primitives also allows us to easily incorporate well-established constraints such as-isometric-as-possible regularisation. While learning this deformation model is prone to local optima when trained in a fully unsupervised manner, we propose to additionally leverage semantic information such as keypoint dynamics to guide the deformation learning. We demonstrate our model with an example application of creating an expressive animatable human avatar from a collection of 3D scans. Here, previous methods mostly rely on variants of the linear blend skinning paradigm, which fundamentally limits the expressivity of such models when dealing with complex cloth appearances such as long skirts. We show the advantages of our dynamic point field framework in terms of its representational power, learning efficiency, and robustness to out-of-distribution novel poses.

cs.CV↗

arXiv2023

CrysFieldExplorer: a software for rapid optimization of crystal field Hamiltonian

We present a new lite python-based program, CrysFieldExplorer, for fast optimizing crystal electric field (CEF) parameters to fit experimental data. The main novelty of CrysFieldExplorer is the development of a unique loss function, referred to as the Spectrum-Characteristic Loss ($L_{\text{Spectrum}}$), which is defined based on the characteristic polynomial of the Hamiltonian matrix. Particle Swarm Optimization and Covariance matrix adaptation evolution strategy are used to find the minimum of the total loss function. We demonstrate that CrysFieldExplorer can performs direct fitting of CEF parameters to any experimental data such as neutron spectrum, susceptibility, magnetizations etc. CrysFieldExplorer can handle a large amount of none-zero CEF parameters and reveal multiple local and global minimum solutions. Detailed crystal field theory, description of the loss function, implementation and limit of the program are discussed within context of two examples.

cond-mat.str-el↗

arXiv2023

Opening the Black Box of the Radiation Belt Machine Learning Model

Many Machine Learning (ML) systems, especially neural networks, are fundamentally regarded as black boxes since it is difficult to grasp how they function once they have been trained. Here, we tackle the issue of the interpretability of a high-accuracy ML model created to model the flux of Earth's radiation belt electrons. The Outer RadIation belt Electron Neural net model (ORIENT) uses only solar wind conditions and geomagnetic indices as input. Using the Deep SHAPley additive explanations (DeepSHAP) method, we show that the `black box' ORIENT model can be successfully explained. Two significant electron flux enhancement events observed by Van Allen Probes during the storm interval of 17 to 18 March 2013 and non storm interval of 19 to 20 September 2013 are investigated using the DeepSHAP method. The results show that the feature importances calculated from the purely data driven ORIENT model identify physically meaningful behavior consistent with current physical understanding.

physics.space-ph↗

arXiv2022

Pair-Based Joint Encoding with Relational Graph Convolutional Networks for Emotion-Cause Pair Extraction

Emotion-cause pair extraction (ECPE) aims to extract emotion clauses and corresponding cause clauses, which have recently received growing attention. Previous methods sequentially encode features with a specified order. They first encode the emotion and cause features for clause extraction and then combine them for pair extraction. This lead to an imbalance in inter-task feature interaction where features extracted later have no direct contact with the former. To address this issue, we propose a novel Pair-Based Joint Encoding (PBJE) network, which generates pairs and clauses features simultaneously in a joint feature encoding manner to model the causal relationship in clauses. PBJE can balance the information flow among emotion clauses, cause clauses and pairs. From a multi-relational perspective, we construct a heterogeneous undirected graph and apply the Relational Graph Convolutional Network (RGCN) to capture the various relationship between clauses and the relationship between pairs and clauses. Experimental results show that PBJE achieves state-of-the-art performance on the Chinese benchmark corpus.

cs.CL↗

arXiv2022

Distilling Causal Effect from Miscellaneous Other-Class for Continual Named Entity Recognition

Continual Learning for Named Entity Recognition (CL-NER) aims to learn a growing number of entity types over time from a stream of data. However, simply learning Other-Class in the same way as new entity types amplifies the catastrophic forgetting and leads to a substantial performance drop. The main cause behind this is that Other-Class samples usually contain old entity types, and the old knowledge in these Other-Class samples is not preserved properly. Thanks to the causal inference, we identify that the forgetting is caused by the missing causal effect from the old data. To this end, we propose a unified causal framework to retrieve the causality from both new entity types and Other-Class. Furthermore, we apply curriculum learning to mitigate the impact of label noise and introduce a self-adaptive weight for balancing the causal effects between new entity types and Other-Class. Experimental results on three benchmark datasets show that our method outperforms the state-of-the-art method by a large margin. Moreover, our method can be combined with the existing state-of-the-art methods to improve the performance in CL-NER

cs.CL↗

arXiv2022

Neural Point-based Shape Modeling of Humans in Challenging Clothing

Parametric 3D body models like SMPL only represent minimally-clothed people and are hard to extend to clothing because they have a fixed mesh topology and resolution. To address these limitations, recent work uses implicit surfaces or point clouds to model clothed bodies. While not limited by topology, such methods still struggle to model clothing that deviates significantly from the body, such as skirts and dresses. This is because they rely on the body to canonicalize the clothed surface by reposing it to a reference shape. Unfortunately, this process is poorly defined when clothing is far from the body. Additionally, they use linear blend skinning to pose the body and the skinning weights are tied to the underlying body parts. In contrast, we model the clothing deformation in a local coordinate space without canonicalization. We also relax the skinning weights to let multiple body parts influence the surface. Specifically, we extend point-based methods with a coarse stage, that replaces canonicalization with a learned pose-independent "coarse shape" that can capture the rough surface geometry of clothing like skirts. We then refine this using a network that infers the linear blend skinning weights and pose dependent displacements from the coarse representation. The approach works well for garments that both conform to, and deviate from, the body. We demonstrate the usefulness of our approach by learning person-specific avatars from examples and then show how they can be animated in new poses and motions. We also show that the method can learn directly from raw scans with missing data, greatly simplifying the process of creating realistic avatars. Code is available for research purposes at {\small\url{https://qianlim.github.io/SkiRT}}.

cs.CV↗

arXiv2022

EgoBody: Human Body Shape and Motion of Interacting People from Head-Mounted Devices

Understanding social interactions from egocentric views is crucial for many applications, ranging from assistive robotics to AR/VR. Key to reasoning about interactions is to understand the body pose and motion of the interaction partner from the egocentric view. However, research in this area is severely hindered by the lack of datasets. Existing datasets are limited in terms of either size, capture/annotation modalities, ground-truth quality, or interaction diversity. We fill this gap by proposing EgoBody, a novel large-scale dataset for human pose, shape and motion estimation from egocentric views, during interactions in complex 3D scenes. We employ Microsoft HoloLens2 headsets to record rich egocentric data streams (including RGB, depth, eye gaze, head and hand tracking). To obtain accurate 3D ground truth, we calibrate the headset with a multi-Kinect rig and fit expressive SMPL-X body meshes to multi-view RGB-D frames, reconstructing 3D human shapes and poses relative to the scene, over time. We collect 125 sequences, spanning diverse interaction scenarios, and propose the first benchmark for 3D full-body pose and shape estimation of the social partner from egocentric views. We extensively evaluate state-of-the-art methods, highlight their limitations in the egocentric scenario, and address such limitations leveraging our high-quality annotations. Data and code are available at https://sanweiliti.github.io/egobody/egobody.html.

cs.CV↗

arXiv2022

Efficient Path Planning in Narrow Passages for Robots with Ellipsoidal Components

Path planning has long been one of the major research areas in robotics, with PRM and RRT being two of the most effective classes of planners. Though generally very efficient, these sampling-based planners can become computationally expensive in the important case of "narrow passages". This paper develops a path planning paradigm specifically formulated for narrow passage problems. The core is based on planning for rigid-body robots encapsulated by unions of ellipsoids. Each environmental feature is represented geometrically using a strictly convex body with a $\mathcal{C}^1$ boundary (e.g., superquadric). The main benefit of doing this is that configuration-space obstacles can be parameterized explicitly in closed form, thereby allowing prior knowledge to be used to avoid sampling infeasible configurations. Then, by characterizing a tight volume bound for multiple ellipsoids, robot transitions involving rotations are guaranteed to be collision-free without needing to perform traditional collision detection. Furthermore, by combining with a stochastic sampling strategy, the proposed planning framework can be extended to solving higher dimensional problems in which the robot has a moving base and articulated appendages. Benchmark results show that the proposed framework often outperforms the sampling-based planners in terms of computational time and success rate in finding a path through narrow corridors for both single-body robots and those with higher dimensional configuration spaces. Physical experiments using the proposed framework are further demonstrated on a humanoid robot that walks in several cluttered environments with narrow passages.

cs.RO↗

arXiv2022

Magnetic Field Tuning of Parallel Spin Stripe Order and Fluctuations near the Pseudogap Quantum Critical Point in La$_{1.36}$Nd$_{0.4}$Sr$_{0.24}$CuO$_4$

A quantum critical point in the single layer, hole-doped cuprate system La$_{1.6-x}$Nd$_{0.4}$Sr$_x$CuO$_4$ (Nd-LSCO), near $x$ = 0.23 has been proposed as an organizing principle for understanding high temperature superconductivity. Our earlier neutron diffraction work on Nd-LSCO at optimal and high doping revealed static parallel spin stripes to exist out to the QCP and slightly beyond, at $x$ = 0.24 and 0.26. We examine more closely the parallel spin stripe order parameter in Nd-LSCO in both zero magnetic field and fields up to 8 T for H // c in these single crystals. In contrast to earlier studies at lower doping, we observe that H //c in excess of $\sim$ 2.5 T eliminates the incommensurate quasi-Bragg peaks associated with parallel spin stripes. But this elastic scattering is not destroyed by the field; rather it is transferred to commensurate {\textbf{Q} = 0} Bragg positions, implying that the spins participating in the spin stripes have been polarized. Inelastic neutron scattering measurements at high fields show an increase in the low energy, parallel spin stripe fluctuations and evidence for a spin gap, $Δ_{spin}$= 3 $\pm$ 0.5 meV for Nd-LSCO with $x$ = 0.24. This is shown to be consistent with spin gap measurements as a function of superconducting T$_C$ over five different families of cuprate superconductors, which follow the approximate linear relation, $Δ_{spin}$ = 3.5 k$_B$T$_C$.

cond-mat.supr-con↗

arXiv2022

Dynamic Parallel Spin Stripes from the 1/8 anomaly to the End of Superconductivity in La$_{1.6-x}$Nd$_{0.4}$Sr$_x$CuO$_4$

We have carried out new neutron spectroscopic measurements on single crystals of La$_{1.6-x}$Nd$_{0.4}$Sr$_x$CuO$_4$ from 0.12 to 0.26 using time-of-flight techniques. These measurements allow us to follow the evolution of parallel spin stripe fluctuations with energies less than 33 meV, from x=0.12 to 0.26. Samples at these hole-doping levels are known to display static (on the neutron scattering time scale) parallel spin stripes at low temperature, with onset temperatures and intensities which decrease rapidly with increasing x. Nonetheless, we report remarkably similar dynamic spectral weight for the corresponding dynamic parallel spin stripes, between 5 meV to 33 meV, from the 1/8 anomaly near x=0.12, to optimal doping near x=0.19 to the quantum critical point for the pseudogap phase near x=0.24, and finally to the approximate end of superconductivity near x=0.26. This observed dynamic magnetic spectral weight is structured in energy with a peak near 17 meV at all dopings studied. Earlier neutron and resonant x-ray scattering measurements on related cuprate superconductors have reported both a disappearance with increasing doping of magnetic fluctuations at ($π$, $π$) wavevectors characterizing parallel spin stripe structures, and persistant paramagnon scattering away from this wavevector, respectively. Our new results on La$_{1.6-x}$Nd$_{0.4}$Sr$_x$CuO$_4$ from 0.12 < x <0.26 clearly show persistent parallel spin stripe fluctuations at and around at ($π$, $π$), and across the full range of doping studied. These results are also compared to recent theory. Together with a rapidly declining x-dependence to the static parallel spin stripe order, the persistent parallel spin stripe fluctuations show a remarkable similarity to the expectations of a quantum spin glass, random t-J model, recently introduced to describe strong local correlations in cuprates.

cond-mat.supr-con↗

arXiv2022

MetaAvatar: Learning Animatable Clothed Human Models from Few Depth Images

In this paper, we aim to create generalizable and controllable neural signed distance fields (SDFs) that represent clothed humans from monocular depth observations. Recent advances in deep learning, especially neural implicit representations, have enabled human shape reconstruction and controllable avatar generation from different sensor inputs. However, to generate realistic cloth deformations from novel input poses, watertight meshes or dense full-body scans are usually needed as inputs. Furthermore, due to the difficulty of effectively modeling pose-dependent cloth deformations for diverse body shapes and cloth types, existing approaches resort to per-subject/cloth-type optimization from scratch, which is computationally expensive. In contrast, we propose an approach that can quickly generate realistic clothed human avatars, represented as controllable neural SDFs, given only monocular depth images. We achieve this by using meta-learning to learn an initialization of a hypernetwork that predicts the parameters of neural SDFs. The hypernetwork is conditioned on human poses and represents a clothed neural avatar that deforms non-rigidly according to the input poses. Meanwhile, it is meta-learned to effectively incorporate priors of diverse body shapes and cloth types and thus can be much faster to fine-tune, compared to models trained from scratch. We qualitatively and quantitatively show that our approach outperforms state-of-the-art approaches that require complete meshes as inputs while our approach requires only depth frames as inputs and runs orders of magnitudes faster. Furthermore, we demonstrate that our meta-learned hypernetwork is very robust, being the first to generate avatars with realistic dynamic cloth deformations given as few as 8 monocular depth frames.

cs.CV↗

arXiv2021

The Power of Points for Modeling Humans in Clothing

Currently it requires an artist to create 3D human avatars with realistic clothing that can move naturally. Despite progress on 3D scanning and modeling of human bodies, there is still no technology that can easily turn a static scan into an animatable avatar. Automating the creation of such avatars would enable many applications in games, social networking, animation, and AR/VR to name a few. The key problem is one of representation. Standard 3D meshes are widely used in modeling the minimally-clothed body but do not readily capture the complex topology of clothing. Recent interest has shifted to implicit surface models for this task but they are computationally heavy and lack compatibility with existing 3D tools. What is needed is a 3D representation that can capture varied topology at high resolution and that can be learned from data. We argue that this representation has been with us all along -- the point cloud. Point clouds have properties of both implicit and explicit representations that we exploit to model 3D garment geometry on a human body. We train a neural network with a novel local clothing geometric feature to represent the shape of different outfits. The network is trained from 3D point clouds of many types of clothing, on many bodies, in many poses, and learns to model pose-dependent clothing deformations. The geometry feature can be optimized to fit a previously unseen scan of a person in clothing, enabling the scan to be reposed realistically. Our model demonstrates superior quantitative and qualitative results in both multi-outfit modeling and unseen outfit animation. The code is available for research purposes.

cs.CV↗

arXiv2021

Relativistic electron flux model in the outer radiation belt using a neural network approach

We present a machine-learning-based model of relativistic electron fluxes >1.8 MeV using a neural network approach in the Earth's outer radiation belt. The Outer RadIation belt Electron Neural net model for Relativistic electrons (ORIENT-R) uses only solar wind conditions and geomagnetic indices as input. For the first time, we show that the state of the outer radiation belt can be determined using only solar wind conditions and geomagnetic indices, without any initial and boundary conditions. The most important features for determining outer radiation belt dynamics are found to be AL, solar wind flow speed and density, and SYM-H indices. ORIENT-R reproduces out-of-sample relativistic electron fluxes with a correlation coefficient of 0.95 and an uncertainty factor of ~2. ORIENT-R reproduces radiation belt dynamics during an out-of-sample geomagnetic storm with good agreement to the observations. In addition, ORIENT-R was run for a completely out-of-sample period between March 2018 and October 2019 when the AL index ended and was replaced with the predicted AL index (lasp.colorado.edu/~lix). It reproduces electron fluxes with a correlation coefficient of 0.92 and an out-of-sample uncertainty factor of ~3. Furthermore, ORIENT-R captured the trend in the electron fluxes from low-earth-orbit (LEO) SAMPEX, which is a completely out-of-sample dataset both temporally and spatially. In sum, the ORIENT-R model can reproduce transport, acceleration, decay, and dropouts of the outer radiation belt anywhere from short timescales (i.e., geomagnetic storms) and very long timescales (i.e., solar cycle) variations.

physics.space-ph↗

arXiv2021

Embedding-based Product Retrieval in Taobao Search

Nowadays, the product search service of e-commerce platforms has become a vital shopping channel in people's life. The retrieval phase of products determines the search system's quality and gradually attracts researchers' attention. Retrieving the most relevant products from a large-scale corpus while preserving personalized user characteristics remains an open question. Recent approaches in this domain have mainly focused on embedding-based retrieval (EBR) systems. However, after a long period of practice on Taobao, we find that the performance of the EBR system is dramatically degraded due to its: (1) low relevance with a given query and (2) discrepancy between the training and inference phases. Therefore, we propose a novel and practical embedding-based product retrieval model, named Multi-Grained Deep Semantic Product Retrieval (MGDSPR). Specifically, we first identify the inconsistency between the training and inference stages, and then use the softmax cross-entropy loss as the training objective, which achieves better performance and faster convergence. Two efficient methods are further proposed to improve retrieval relevance, including smoothing noisy training data and generating relevance-improving hard negative samples without requiring extra knowledge and training procedures. We evaluate MGDSPR on Taobao Product Search with significant metrics gains observed in offline experiments and online A/B tests. MGDSPR has been successfully deployed to the existing multi-channel retrieval system in Taobao Search. We also introduce the online deployment scheme and share practical lessons of our retrieval system to contribute to the community.

cs.IR↗