BZPEERReading, trust and collaboration for research
Menu

Discover

GraphFollow connections around a paper, topic or saved view.

Workspaces

HomePick up where your research flow left off.InboxHandle requests, replies and notifications from one place.CollectionsCurate reading lists, evidence sets and shareable dossiers.ResearchSee your private provenance graph, trail and imports.CollaborationMove from discovery into focused discussion rooms.PublishDraft, preview and publish full-text research articles.

Network

ExpertsFind specialists worth following or asking for help.CommunitiesJoin research circles organized around shared work.PartnersSee external organizations active around the same topics.

Opportunities

ListingsBrowse roles, calls and collaborations with clearer fit signals.

Account

Log inReturn to your reading and collaboration workspace.Create accountStart saving work, following people and joining teams.
Log inCreate account

Researcher profile

Chi Chen

Chi Chen contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate
cond-mat.mtrl-sciphysics.comp-phComputation and LanguageComputer Visioncond-mat.dis-nncond-mat.mes-hallphysics.opticsq-fin.ST

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author
12works
0followers
8topics
4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Identity and collaboration

How to connect with this researcher

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Log in to claim

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Research graph

See the researcher in context

Open full explorer

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Published work

12 published item(s)

preprint2026arXiv

MiniCPM-o 4.5: Towards Real-Time Full-Duplex Omni-Modal Interaction

Recent progress in multimodal large language models (MLLMs) has brought AI capabilities from static offline data processing to real-time streaming interaction, yet they still remain far from human-level multimodal interaction. The key bottlenecks are no longer modality coverage or latency alone, but the interaction paradigm itself. First, perception and response are still separated into alternating phases, preventing models from incorporating new inputs for timely adjustment during generation. Second, most current models remain reactive, responding only to explicit user requests instead of acting proactively in the evolving multimodal environment. We present MiniCPM-o 4.5, our latest effort towards human-like multimodal interaction, which mitigates these gaps by real-time full-duplex omni-modal interaction. It can see, listen, and speak simultaneously in real-time, while also exhibiting proactive behaviors such as issuing reminders or comments based on its continuous understanding of the live scene. The key technique behind MiniCPM-o 4.5 is Omni-Flow, a unified streaming framework that aligns omni-modal inputs and outputs along a shared temporal axis. This formulation converts conventional turn-based interaction into a full-duplex, time-aligned process, enabling simultaneous perception and response and allowing proactive behavior to arise within the same framework. With a total of 9B parameters, MiniCPM-o 4.5 approaches Gemini 2.5 Flash in vision-language capabilities, delivering state-of-the-art open-source performance at its scale. It also surpasses Qwen3-Omni-30B-A3B in omni-modal understanding and delivers better speech generation, with significantly higher computation efficiency. Driven by its efficient architecture design and inference optimization, the model can perform real-time full-duplex omni-modal interaction on edge devices with less than 12GB RAM cost.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

A Universal Machine Learning Model for Elemental Grain Boundary Energies

The grain boundary (GB) energy has a profound influence on the grain growth and properties of polycrystalline metals. Here, we show that the energy of a GB, normalized by the bulk cohesive energy, can be described purely by four geometric features. By machine learning on a large computed database of 361 small $Σ$ ($Σ< 10$) GBs of more than 50 metals, we develop a model that can predict the grain boundary energies to within a mean absolute error of 0.13 J m$^{-2}$. More importantly, this universal GB energy model can be extrapolated to the energies of high $Σ$ GBs without loss in accuracy. These results highlight the importance of capturing fundamental scaling physics and domain knowledge in the design of interpretable, extrapolatable machine learning models for materials science.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

CG-SSD: Corner Guided Single Stage 3D Object Detection from LiDAR Point Cloud

At present, the anchor-based or anchor-free models that use LiDAR point clouds for 3D object detection use the center assigner strategy to infer the 3D bounding boxes. However, in a real world scene, the LiDAR can only acquire a limited object surface point clouds, but the center point of the object does not exist. Obtaining the object by aggregating the incomplete surface point clouds will bring a loss of accuracy in direction and dimension estimation. To address this problem, we propose a corner-guided anchor-free single-stage 3D object detection model (CG-SSD ).Firstly, 3D sparse convolution backbone network composed of residual layers and sub-manifold sparse convolutional layers are used to construct bird&#39;s eye view (BEV) features for further deeper feature mining by a lite U-shaped network; Secondly, a novel corner-guided auxiliary module (CGAM) is proposed to incorporate corner supervision signals into the neural network. CGAM is explicitly designed and trained to detect partially visible and invisible corners to obtains a more accurate object feature representation, especially for small or partial occluded objects; Finally, the deep features from both the backbone networks and CGAM module are concatenated and fed into the head module to predict the classification and 3D bounding boxes of the objects in the scene. The experiments demonstrate CG-SSD achieves the state-of-art performance on the ONCE benchmark for supervised 3D object detection using single frame point cloud data, with 62.77%mAP. Additionally, the experiments on ONCE and Waymo Open Dataset show that CGAM can be extended to most anchor-based models which use the BEV feature to detect objects, as a plug-in and bring +1.17%-+14.27%AP improvement.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

The Intercalation Chemistry of the Disordered RockSalt Li3V2O5 Anode from Cluster Expansions and Machine Learning Interatomic Potentials

Disordered rocksalt (DRX) Li3V2O5 is a promising candidate for anode in rechargeable lithium-ion batteries because of its ideal low voltage, high rate capability, and superior cycling stability. Herein, we presents a comprehensive study of intercalation chemistry of the DRX-Li3V2O5 anode using density functional theory calculations combined with machine learning cluster expansions and interatomic potentials. The predicted voltage profile of the disordered Li3V2O5 anode at room temperature based on Monte Carlo simulations with a fitted cluster expansion model is in excellent agreement with experiments. In contrast to previous DFT results, we find that Li ions predominately intercalate into tetrahedral sites during charging, while the majority of Li and V ions at octahedral sites remain stable. In addition, MD simulations with a fitted moment tensor potential attribute the fast-charging capability of DRX-Li3V2O5 to the facile diffusivity of Li+ via tetrahedral - octahedral - tetrahedral pathway. We further suggest tuning the Li:V ratio as a means to trade off increased lithiation capacity and decreased anode voltage in this system. This work provides in-depth insights into the high-performance DRX-Li3V2O5 anode, and paves the way to the discovery of other disordered anode materials.

0 citations0 reviewsNo code linkedOpen access
preprint2022arXiv

Visualization of Band Shifting and Interlayer Coupling in WxMo1-xS2 Alloys using Near-Field Broadband Absorption Microscopy

Beyond-diffraction-limit optical absorption spectroscopy provides profound information on the graded band structures of composition-spread and stacked two-dimensional materials, in which direct/indirect bandgap, interlayer coupling, sliding, and possible defects significantly modify their optoelectronic functionalities such as photoluminescence efficiency. We here visualize the spatially-varying band structure of monolayer and bilayer transition metal dichalcogenide alloys for the first time by using near-field broadband absorption microscopy. The near-field-spectral and -spatial diagrams manifest the excitonic band shift that results from the interplay of composition spreading and interlayer coupling. These results enable us to identify the top layer of the bilayer alloy as pure WS2. We also use the aberration-free near-field transmittance images to demarcate the exact boundaries of alloyed and pure transition metal dichalcogenides. This technology can offer new insights on various layered structures in the era of stacking science in quest of novel quantum optoelectronic devices.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Learning Properties of Ordered and Disordered Materials from Multi-fidelity Data

Predicting the properties of a material from the arrangement of its atoms is a fundamental goal in materials science. While machine learning has emerged in recent years as a new paradigm to provide rapid predictions of materials properties, their practical utility is limited by the scarcity of high-fidelity data. Here, we develop multi-fidelity graph networks as a universal approach to achieve accurate predictions of materials properties with small data sizes. As a proof of concept, we show that the inclusion of low-fidelity Perdew-Burke-Ernzerhof band gaps greatly enhances the resolution of latent structural features in materials graphs, leading to a 22-45\% decrease in the mean absolute errors of experimental band gap predictions. We further demonstrate that learned elemental embeddings in materials graph networks provide a natural approach to model disorder in materials, addressing a fundamental gap in the computational prediction of materials properties.

0 citations0 reviewsNo code linkedOpen access
preprint2021arXiv

Recent Advances and Applications of Deep Learning Methods in Materials Science

Deep learning (DL) is one of the fastest growing topics in materials data science, with rapidly emerging applications spanning atomistic, image-based, spectral, and textual data modalities. DL allows analysis of unstructured data and automated identification of features. Recent development of large materials databases has fueled the application of DL methods in atomistic prediction in particular. In contrast, advances in image and spectral data have largely leveraged synthetic data enabled by high quality forward models as well as by generative unsupervised DL methods. In this article, we present a high-level overview of deep-learning methods followed by a detailed discussion of recent developments of deep learning in atomistic simulation, materials imaging, spectral analysis, and natural language processing. For each modality we discuss applications involving both theoretical and experimental data, typical modeling approaches with their strengths and limitations, and relevant publicly available software and datasets. We conclude the review with a discussion of recent cross-cutting work related to uncertainty quantification in this field and a brief perspective on limitations, challenges, and potential growth areas for DL methods in materials science. The application of DL methods in materials science presents an exciting avenue for future materials discovery and design.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Complex Strengthening Mechanisms in the NbMoTaW Multi-Principal Element Alloy

Refractory multi-principal element alloys (MPEAs) have exceptional mechanical properties, including high strength-to-weight ratio and fracture toughness, at high temperatures. Here, we elucidate the complex interplay between segregation, short range order and strengthening in the NbMoTaW MPEA through atomistic simulations with a highly accurate machine learning interatomic potential. In the single crystal MPEA, we find greatly reduced anisotropy in the critically resolved shear stress between screw and edge dislocations compared to the elemental metals. In the polycrystalline MPEA, we demonstrate that thermodynamically-driven Nb segregation to the grain boundaries (GBs) and W enrichment within the grains intensifies the observed short range order (SRO). The increased GB stability due to Nb enrichment reduces the von Mises strain, resulting in higher strength than a random solid-solution MPEA. These results highlight the need to simultaneously tune GB composition and bulk SRO to tailor the mechanical properties of MPEAs.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Genetic Algorithm-Guided Deep Learning of Grain Boundary Diagrams: Addressing the Challenge of Five Degrees of Freedom

Grain boundaries (GBs) often control the processing and properties of polycrystalline materials. Here, a potentially transformative research is represented by constructing GB property diagrams as functions of temperature and bulk composition, also called &#34;complexion diagrams,&#34; as a general materials science tool on par with phase diagrams. However, a GB has five macroscopic (crystallographic) degrees of freedom (DOFs). It is essentially a &#34;mission impossible&#34; to construct property diagrams for GBs as a function of five DOFs by either experiments or modeling. Herein, we combine isobaric semi-grand-canonical ensemble hybrid Monte Carlo and molecular dynamics (hybrid MC/MD) simulations with a genetic algorithm (GA) and deep neural network (DNN) models to tackle this grand challenge. The DNN prediction is ~108 faster than atomistic simulations, thereby enabling the construction of the property diagrams for millions of distinctly different GBs of five DOFs. Notably, excellent prediction accuracies have been achieved for not only symmetric-tilt and twist GBs, but also asymmetric-tilt and mixed tilt-twist GBs; the latter are more complex and much less understood, but they are ubiquitous and often limit the performance properties of real polycrystals as the weak links. The data-driven prediction of GB properties as function of temperature, bulk composition, and five crystallographic DOFs (i.e., in a 7D space) opens a new paradigm.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Trimming the Sail: A Second-order Learning Paradigm for Stock Prediction

Nowadays, machine learning methods have been widely used in stock prediction. Traditional approaches assume an identical data distribution, under which a learned model on the training data is fixed and applied directly in the test data. Although such assumption has made traditional machine learning techniques succeed in many real-world tasks, the highly dynamic nature of the stock market invalidates the strict assumption in stock prediction. To address this challenge, we propose the second-order identical distribution assumption, where the data distribution is assumed to be fluctuating over time with certain patterns. Based on such assumption, we develop a second-order learning paradigm with multi-scale patterns. Extensive experiments on real-world Chinese stock data demonstrate the effectiveness of our second-order learning paradigm in stock prediction.

0 citations0 reviewsNo code linkedOpen access
preprint2020arXiv

Unified Theory of Thermal Quenching in Inorganic Phosphors

We unify two prevailing theories of thermal quenching (TQ) in rare-earth-activated inorganic phosphors - the cross-over and auto-ionization mechanisms - into a single predictive model. Crucially, we have developed computable descriptors for activator environment stability from ab initio molecular dynamics simulations to predict TQ under the cross-over mechanism, which can be augmented by a band gap calculation to account for auto-ionization. The resulting TQ model predicts the experimental TQ in 29 known phosphors to within ~ 3-8%. Finally, we have developed an efficient topological approach to rapidly screen vast chemical spaces for the discovery of novel, thermally robust phosphors.

0 citations0 reviewsNo code linkedOpen access
preprint2019arXiv

A Performance and Cost Assessment of Machine Learning Interatomic Potentials

Machine learning of the quantitative relationship between local environment descriptors and the potential energy surface of a system of atoms has emerged as a new frontier in the development of interatomic potentials (IAPs). Here, we present a comprehensive evaluation of ML-IAPs based on four local environment descriptors --- Behler-Parrinello symmetry functions, smooth overlap of atomic positions (SOAP), the Spectral Neighbor Analysis Potential (SNAP) bispectrum components, and moment tensors --- using a diverse data set generated using high-throughput density functional theory (DFT) calculations. The data set comprising bcc (Li, Mo) and fcc (Cu, Ni) metals and diamond group IV semiconductors (Si, Ge) is chosen to span a range of crystal structures and bonding. All descriptors studied show excellent performance in predicting energies and forces far surpassing that of classical IAPs, as well as predicting properties such as elastic constants and phonon dispersion curves. We observe a general trade-off between accuracy and the degrees of freedom of each model, and consequently computational cost. We will discuss these trade-offs in the context of model selection for molecular dynamics and other applications.

0 citations0 reviewsNo code linkedOpen access