papers

Publications (74)

cs.LG2023

From block-Toeplitz matrices to differential equations on graphs: towards a general theory for scalable masked Transformers

Krzysztof Choromanski, Han Lin, Haoxian Chen +7

In this paper we provide, to the best of our knowledge, the first comprehensive approach for incorporating various masking mechanisms into Transformers architectures in a scalable…

cs.CV2026

EPiC: Efficient Video Camera Control Learning with Precise Anchor-Video Guidance

Zun Wang, Jaemin Cho, Jialu Li +4

Recent approaches for video generation with camera control often create anchor videos (i.e., rendered videos that approximate desired camera motions) to guide diffusion models as a…

cs.CV2024

VideoDirectorGPT: Consistent Multi-scene Video Generation via LLM-Guided Planning

Han Lin, Abhay Zala, Jaemin Cho +1

Recent text-to-video (T2V) generation methods have seen significant advancements. However, the majority of these works focus on producing short video clips of a single event (i.e.,…

physics.flu-dyn2022

FSVPy: A Python-based Package for Fluorescent Streak Velocimetry (FSV)

Han Lin, Brendan C. Blackwell, Connor C. Call +4

Predictive constitutive equations that connect easy-to-measure transport properties (e.g., viscosity and conductivity) with system performance variables (e.g., power consumption an…

astro-ph.HE2025

Supernovae at Distances < 40 Mpc: I.Catalogues and fractions of Supernovae in a Complete Sample

Xiaoran Ma, Xiaofeng Wang, Jun Mo +20

Context.This is the first paper of a series aiming to determine the fractions and birth rates of various types of supernovae (SNe) in the local Universe. Aims. In this paper, we ai…

astro-ph.HE2024

Probing the Shock Breakout Signal of SN 2024ggi from the Transformation of Early Flash Spectroscopy

Jujia Zhang, Luc Dessart, Xiaofeng Wang +11

We present early-time, hour-to-day cadence spectroscopy of the nearby type II supernova (SN II) 2024ggi, which was discovered at a phase when the SN shock just emerged from the red…

astro-ph.IM2020

The Tsinghua University-Ma Huateng Telescopes for Survey: Overview and Performance of the System

Ji-Cheng Zhang, Xiao-Feng Wang, Jun Mo +16

Over the past decade, time-domain astronomy in optical bands has developed rapidly with the operations of some wide-field survey facilities. However, most of these surveys are cond…

astro-ph.HE2023

A Superluminous Supernova Lightened by Collisions with Pulsational Pair-instability Shells

Weili Lin, Xiaofeng Wang, Lin Yan +23

Superluminous supernovae are among the most energetic stellar explosions in the Universe, but their energy sources remain an open question. Here we present long-term observations o…

cs.LG2023

Efficient Graph Field Integrators Meet Point Clouds

Krzysztof Choromanski, Arijit Sehanobish, Han Lin +13

We present two new classes of algorithms for efficient field integration on graphs encoding point clouds. The first class, SeparatorFactorization(SF), leverages the bounded genus o…

astro-ph.HE2026

Pre-Supernova Eruptions Triggered by Sudden Energy Deposition in Low-Mass Core-Collapse Supernova Progenitors

Shuai Zha, Han Lin, Xuefei Chen +1

In low-mass core-collapse supernova (CCSN) progenitors, nuclear burning beyond oxygen can become explosive under degenerate conditions, triggering eruptive mass loss before the fin…

astro-ph.HE2024

Probing diversity of type II supernovae with the Chinese Space Station Telescope

Han Lin, Jujia Zhang, Xinghan Zhang

Type II supernovae (SNe II), which show abundant hydrogen in their spectra, belong to a class of SNe with diverse observed properties. It is commonly accepted that SNe II are produ…

cs.CV2025

Exploring MLLM-Diffusion Information Transfer with MetaCanvas

Han Lin, Xichen Pan, Ziqi Huang +10

Multimodal learning has rapidly advanced visual understanding, largely via multimodal large language models (MLLMs) that use powerful LLMs as cognitive cores. In visual generation,…

astro-ph.HE2019

SN 2017gmr: An energetic Type II-P supernova with asymmetries

Jennifer E. Andrews, D. J. Sand, S. Valenti +77

We present high-cadence ultraviolet (UV), optical, and near-infrared (NIR) data on the luminous Type II-P supernova SN 2017gmr from hours after discovery through the first 180 days…

cs.LG2026

Humanity's Last Exam

Long Phan, Alice Gatti, Ziwen Han +1144

Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achi…

astro-ph.HE2021

Optical and spectral observations and hydrodynamic modelling of Type IIb Supernova 2017gpn

Elena A. Balakina, Maria V. Pruzhinskaya, Alexander S. Moskvitin +6

In this work we present the photometric and spectroscopic observations of Type IIb Supernova 2017gpn. This supernova was discovered in the error-box of LIGO/Virgo G299232 gravitati…

cs.CL2026

JoyAI-LLM Flash: Advancing Mid-Scale LLMs with Token Efficiency

Aichen Cai, Anmeng Zhang, Anyu Li +66

We introduce JoyAI-LLM Flash, an efficient Mixture-of-Experts (MoE) language model designed to redefine the trade-off between strong performance and token efficiency in the sub-50B…

physics.optics2025

Deep-learning-enabled inverse design of large-scale metasurfaces with full-wave accuracy

Borui Xu, Jingzhu Shao, Xiangyu Zhao +8

Recent advances in meta-optics have enabled diverse functionalities in compact optical devices; however, conventional forward design approaches become inadequate as device complexi…

astro-ph.HE2021

SN 2017hpa: A Nearby Carbon-Rich Type Ia Supernova with a Large Velocity Gradient

Xiangyun Zeng, Xiaofeng Wang, Ali Esamdin +24

We present extensive, well-sampled optical and ultraviolet photometry and optical spectra of the Type Ia supernova (SN Ia) 2017hpa. The light curves indicate that SN 2017hpa is a n…

cs.CV2025

Training-free Guidance in Text-to-Video Generation via Multimodal Planning and Structured Noise Initialization

Jialu Li, Shoubin Yu, Han Lin +3

Recent advancements in text-to-video (T2V) diffusion models have significantly enhanced the visual quality of the generated videos. However, even recent T2V models find it challeng…

hep-th2025

D-commuting SYK model: building quantum chaos from integrable blocks

Ping Gao, Han Lin, Cheng Peng

We construct a new family of quantum chaotic models by combining multiple copies of integrable commuting SYK models. As each copy of the commuting SYK model does not commute with o…

cs.CV2026

V-Co: A Closer Look at Visual Representation Alignment via Co-Denoising

Han Lin, Xichen Pan, Zun Wang +4

Pixel-space diffusion has recently re-emerged as a strong alternative to latent diffusion, enabling high-quality generation without pretrained autoencoders. However, standard pixel…

cs.CV2023

Self-supervised Learning for Segmentation and Quantification of Dopamine Neurons in Parkinson's Disease

Fatemeh Haghighi, Soumitra Ghosh, Hai Ngu +5

Parkinson's Disease (PD) is the second most common neurodegenerative disease in humans. PD is characterized by the gradual loss of dopaminergic neurons in the Substantia Nigra (SN)…

astro-ph.SR2018

SN 2014J in M82: New Insights On the Spectral Diversity of Type Ia Supernovae

Kaicheng Zhang, Xiaofeng Wang, JuJia Zhang +25

We present extensive spectroscopic observations for one of the closest type Ia supernovae (SNe Ia), SN 2014J discovered in M82, ranging from 10.4 days before to 473.2 days after B-…

cs.LG2020

Demystifying Orthogonal Monte Carlo and Beyond

Han Lin, Haoxian Chen, Tianyi Zhang +2

Orthogonal Monte Carlo (OMC) is a very effective sampling algorithm imposing structural geometric conditions (orthogonality) on samples for variance reduction. Due to its simplicit…

astro-ph.HE2026

Spectral Dataset of Stripped-Envelope Supernovae from the Tsinghua Supernova Group

Danfeng Xiang, Xiaofeng Wang, Jujia Zhang +40

The extent of envelope stripping in the progenitor stars is directly reflected in the diversity of spectral features observed in stripped-envelope supernovae (SESNe). Through exten…

astro-ph.HE2024

SN 2021dbg: A Luminous Type IIP-IIL Supernova Exploding from a Massive Star with a Layered Shell

Zeyi Zhao, Jujia Zhang, Liping Li +9

We present extensive observations and analysis of supernova (SN) 2021dbg, utilizing optical photometry and spectroscopy. For approximately 385 days following the explosion, SN 2021…

cs.CV2025

Planning with Sketch-Guided Verification for Physics-Aware Video Generation

Yidong Huang, Zun Wang, Han Lin +5

Recent video generation approaches increasingly rely on planning intermediate control signals such as object trajectories to improve temporal coherence and motion fidelity. However…

cs.LG2026

OPD+: Rethinking the Advantage Design for On-Policy Distillation

Hanyang Zhao, Haoxian Chen, Han Lin +3

On-policy distillation (OPD) is a widely used technique to transfer capabilities from capable teacher language models to the base student models, and can be formulated in a reinfor…

astro-ph.HE2024

A spectral data release for 104 Type II Supernovae from the Tsinghua Supernova Group

Han Lin, Xiaofeng Wang, Jujia Zhang +33

We present 206 unpublished optical spectra of 104 type II supernovae obtained by the Xinglong 2.16m telescope and Lijiang 2.4m telescope during the period from 2011 to 2018, spanni…

cs.CV2024

MMT-Bench: A Comprehensive Multimodal Benchmark for Evaluating Large Vision-Language Models Towards Multitask AGI

Kaining Ying, Fanqing Meng, Jin Wang +19

Large Vision-Language Models (LVLMs) show significant strides in general-purpose multimodal applications such as visual dialogue and embodied navigation. However, existing multimod…

cs.LG2026

E^2-LLM: Bridging Neural Signals and Interpretable Affective Analysis

Fei Ma, Han Lin, Yifan Xie +4

Emotion recognition from electroencephalography (EEG) signals remains challenging due to high inter-subject variability, limited labeled data, and the lack of interpretable reasoni…

cs.CV2024

DiagrammerGPT: Generating Open-Domain, Open-Platform Diagrams via LLM Planning

Abhay Zala, Han Lin, Jaemin Cho +1

Text-to-image (T2I) generation has seen significant growth over the past few years. Despite this, there has been little work on generating diagrams with T2I models. A diagram is a…

astro-ph.HE2021

SN 2017fgc: A Fast-Expanding Type Ia Supernova Exploded in Massive Shell Galaxy NGC 474

Xiangyun Zeng, Xiaofeng Wang, Ali Esamdin +31

We present extensive optical photometric and spectroscopic observations of the high-velocity (HV) Type Ia supernova (SN Ia) 2017fgc, covering the phase from 12 d before to $…

astro-ph.HE2025

Supernovae at Distances < 40 Mpc: II. Supernova Rate in the Local Universe

Xiaoran Ma, Xiaofeng Wang, Jun Mo +21

Context.This is the second paper of a series aiming to determine the birth rates of supernovae in the local Universe. Aims. In this paper, we aim to estimate the SN rates in the lo…

cs.CV2024

VEDIT: Latent Prediction Architecture For Procedural Video Representation Learning

Han Lin, Tushar Nagarajan, Nicolas Ballas +4

Procedural video representation learning is an active research area where the objective is to learn an agent which can anticipate and forecast the future given the present video in…

astro-ph.SR2022

Multiwavelength observations of Swift J0243.6+6124 from 2017 to 2022

Wei Liu, Jingzhi Yan, Pablo Reig +8

We have obtained optical spectroscopy and photometry data during four years after the event. The long-term photometric light-curve and the equivalent widths of the Halpha and He I…

econ.GN2026

Unveiling contrasting impacts of heat mitigation and adaptation policies on U.S. internal migration

Chao Li, Xing Su, Chao Fan +7

While climate-induced population migration has received rising attention, the role played by human climate endeavors remains underexplored. Here, we combine machine learning with a…

cs.LG2024

Fast Tree-Field Integrators: From Low Displacement Rank to Topological Transformers

Krzysztof Choromanski, Arijit Sehanobish, Somnath Basu Roy Chowdhury +4

We present a new class of fast polylog-linear algorithms based on the theory of structured matrices (in particular low displacement rank) for integrating tensor fields defined on w…

astro-ph.SR2023

Minute-cadence Observations of the LAMOST Fields with the TMTS: III. Statistic Study of the Flare Stars from the First Two Years

Qichun Liu, Jie Lin, Xiaofeng Wang +22

Tsinghua University-Ma Huateng Telescopes for Survey (TMTS) aims to detect fast-evolving transients in the Universe, which has led to the discovery of thousands of short-period var…

astro-ph.HE2021

The Peculiar Transient AT2018cow: A Possible Origin of A Type Ibn/IIn Supernova

Danfeng Xiang, Xiaofeng Wang, Weili Lin +47

We present our photometric and spectroscopic observations on the peculiar transient AT2018cow. The multi-band photometry covers from peak to 70 days and the spectroscopy rang…

cs.CV2025

Bifrost-1: Bridging Multimodal LLMs and Diffusion Models with Patch-level CLIP Latents

Han Lin, Jaemin Cho, Amir Zadeh +2

There is growing interest in integrating high-fidelity visual synthesis capabilities into large language models (LLMs) without compromising their strong reasoning capabilities. Exi…

physics.optics2022

Design of wavelength division multiplexing devices based on tunable edge states of valley photonic crystals

YuHui Han, HongMing Fei, Han Lin +6

Wavelength division multiplexing (WDM) devices are key elements of Photonic integrated circuits (PICs). Conventional WDM devices based on silicon waveguides and photonic crystals h…

astro-ph.SR2023

The Dusty and Extremely Red Progenitor of the Type II Supernova 2023ixf in Messier 101

Danfeng Xiang, Jun Mo, Lingzhi Wang +4

Stars with initial masses in the range of 8-25 solar masses are thought to end their lives as hydrogen-rich supernovae (SNe II). Based on the pre-explosion images of Hubble Space T…

astro-ph.HE2021

SN 2015bf: a fast declining type II supernova with flash-ionised signatures

Han Lin, Xiaofeng Wang, Jujia Zhang +19

We present optical and ultraviolet photometry, as well as optical spectra, for the type II supernova (SN) 2015bf. Our observations cover the phases from to d af…

cs.CL2024

EnvGen: Generating and Adapting Environments via LLMs for Training Embodied Agents

Abhay Zala, Jaemin Cho, Han Lin +2

Recent SOTA approaches for embodied learning via interaction directly employ large language models (LLMs) as agents to determine the next steps in an environment. Due to their worl…

physics.optics2022

High-performance chiral all-optical logic gate based on topological edge states of valley photonic crystal

Xiaorong Wang, Hongming Fei, Han Lin +6

For all-optical communication and information processing, it is necessary to develop all-optical logic gates based on photonic structures that can directly perform logic operations…

astro-ph.HE2022

SN 2019va: A Type IIP Supernova with Large Influence of Nickel-56 Decay on the Plateau-phase Light Curve

Xinghan Zhang, Xiaofeng Wang, Hanna Sai +12

We present multi-band photometric and spectroscopic observations of the type II supernova, (SN) 2019va, which shows an unusually flat plateau-phase evolution in its V-band light cu…

astro-ph.IM2021

Minute-cadence Observations of the LAMOST Fields with the TMTS: I. Methodology of Detecting Short-period Variables and Results from the first-year Survey

Jie Lin, Xiaofeng Wang, Jun Mo +18

Tsinghua University-Ma Huateng Telescopes for Survey (TMTS), located at Xinglong Station of NAOC, has a field of view upto 18 deg^2. The TMTS has started to monitor the LAMOST sky…

cs.LG2022

Hybrid Random Features

Krzysztof Choromanski, Haoxian Chen, Han Lin +10

We propose a new class of random feature methods for linearizing softmax and Gaussian kernels called hybrid random features (HRFs) that automatically adapt the quality of kernel es…

cs.LG2026

Hinge Regression Tree: A Newton Method for Oblique Regression Tree Splitting

Hongyi Li, Han Lin, Jun Xu

Oblique decision trees combine the transparency of trees with the power of multivariate decision boundaries, but learning high-quality oblique splits is NP-hard, and practical meth…

cs.CV2023

Supervised Masked Knowledge Distillation for Few-Shot Transformers

Han Lin, Guangxing Han, Jiawei Ma +3

Vision Transformers (ViTs) emerge to achieve impressive performance on many data-abundant computer vision tasks by capturing long-range dependencies among local features. However,…

cs.CV2024

Ctrl-Adapter: An Efficient and Versatile Framework for Adapting Diverse Controls to Any Diffusion Model

Han Lin, Jaemin Cho, Abhay Zala +1

ControlNets are widely used for adding spatial control to text-to-image diffusion models with different conditions, such as depth maps, scribbles/sketches, and human poses. However…

cs.CV2024

MCM: Multi-condition Motion Synthesis Framework

Zeyu Ling, Bo Han, Yongkang Wongkan +3

Conditional human motion synthesis (HMS) aims to generate human motion sequences that conform to specific conditions. Text and audio represent the two predominant modalities employ…

cs.CV2026

PhyMotion: Structured 3D Motion Reward for Physics-Grounded Human Video Generation

Yidong Huang, Zun Wang, Han Lin +6

Generating realistic human motion is a central yet unsolved challenge in video generation. While reinforcement learning (RL)-based post-training has driven recent gains in general…

astro-ph.HE2024

Newly Formed Dust within the Circumstellar Environment of SNIa-CSM 2018evt

Lingzhi Wang, Maokai Hu, Lifan Wang +44

Dust associated with various stellar sources in galaxies at all cosmic epochs remains a controversial topic, particularly whether supernovae (SNe) play an important role in dust pr…

astro-ph.SR2020

Can Helium-detonation Model Explain the Observed Diversity of Type Ia Supernovae?

Wenxiong Li, Xiaofeng Wang, Mattia Bulla +13

We study a sample of 16 Type Ia supernovae (SNe Ia) having both spectroscopic and photometric observations within 2 3 days after the first light. The early colors of such…

cs.CV2025

DreamRunner: Fine-Grained Compositional Story-to-Video Generation with Retrieval-Augmented Motion Adaptation

Zun Wang, Jialu Li, Han Lin +2

Storytelling video generation (SVG) aims to produce coherent and visually rich multi-scene videos that follow a structured narrative. Existing methods primarily employ LLM for high…

cs.CV2023

TANDEM3D: Active Tactile Exploration for 3D Object Recognition

Jingxi Xu, Han Lin, Shuran Song +1

Tactile recognition of 3D objects remains a challenging task. Compared to 2D shapes, the complex geometry of 3D surfaces requires richer tactile signals, more dexterous actions, an…

cs.HC2026

Remember Me, Not Save Me: A Collective Memory System for Evolving Virtual Identities in Augmented Reality

Tongzhou Yu, Han Lin

This paper presents "Remember Me, Not Save Me," an AR & AI system enabling virtual citizens to develop personality through collective dialogue. Core innovations include: Dynamic Co…

astro-ph.HE2024

The Red Supergiant Progenitor of Type II Supernova 2024ggi

Danfeng Xiang, Jun Mo, Xiaofeng Wang +8

We present a detailed analysis of the progenitor and its local environment for the recently discovered type II supernova (SN) 2024ggi at a distance of about 6.7~Mpc, by utilizing t…

nlin.CD2021

Theoretical and Experimental Investigation into the Branched Flow Phenomenon of Light

Han Lin, Xiaoyue Ma, Kefan Wang

Branched flow can be observed when a laser beam is coupled into a soap film. This research theoretically explored the phenomenon through analogy between light wave and particles in…

cs.CV2026

Seeing Isn't Knowing: Do VLMs Know When Not to Answer Spatial Questions (and Why)?

Yue Zhang, Zun Wang, Han Lin +3

Spatial reasoning is a fundamental capability for vision-language models (VLMs) deployed in real-world environments. However, visual observations are inherently limited representat…

physics.optics2022

On-chip ultra-compact hexagonal boron nitride topological ring-resonator in visible region

Min Wu, Yibiao Yang, Hongming Fei +3

Ultra-compact topological ring-resonators with chirality are important devices for quantum optics. However, there are limited demonstrations of chiral resonators, especially in the…

astro-ph.SR2022

An 18.9-minute Blue Large-Amplitude Pulsator Crossing the 'Hertzsprung Gap' of Hot Subdwarfs

Jie Lin, Chengyuan Wu, Xiaofeng Wang +28

Blue large-amplitude pulsators (BLAPs) represent a new and rare class of hot pulsating stars with unusually large amplitudes and short periods. Up to now, only 24 confirmed BLAPs h…

cs.CV2025

VideoTG-R1: Boosting Video Temporal Grounding via Curriculum Reinforcement Learning on Reflected Boundary Annotations

Lu Dong, Haiyu Zhang, Han Lin +8

Video temporal grounding (VTG) aims to locate precise segments in videos based on language queries, which is a fundamental challenge in video understanding. While recent Multimodal…

cs.CV2025

Error-Driven Scene Editing for 3D Grounding in Large Language Models

Yue Zhang, Zun Wang, Han Lin +5

Despite recent progress in 3D-LLMs, they remain limited in accurately grounding language to visual and spatial elements in 3D environments. This limitation stems in part from train…

astro-ph.HE2023

Circumstellar Material Ejected Violently by A Massive Star Immediately before its Death

Jujia Zhang, Han Lin, Xiaofeng Wang +7

Type II supernovae represent the most common stellar explosions in the Universe, for which the final stage evolution of their hydrogen-rich massive progenitors towards core-collaps…

astro-ph.HE2021

Optical and Ultraviolet Monitoring of the Black Hole X-ray Binary MAXI J1820+070/ASASSN-18ey for 18 Months

Hanna Sai, Xiaofeng Wang, Jianfeng Wu +20

MAXI J1820+070 is a low-mass black hole X-ray binary system with high luminosity in both optical and X-ray bands during the outburst periods. We present extensive photometry in X-r…

physics.optics2025

Fully automatic fabrication of fibre Bragg gratings using an AI-powered femtosecond laser inscription system

Wenbo Liu, Guiyuan Cao, Zian Liu +6

Fibre Bragg gratings (FBGs) are widely used in optical sensing and communication systems. Femtosecond laser inscription (FLI) enables hydrogen-free, thermally stable, high-resoluti…

cs.CV2026

AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories

Zun Wang, Han Lin, Jaehong Yoon +3

Maintaining spatial world consistency over long horizons remains a central challenge for camera-controllable video generation. Existing memory-based approaches often condition gene…

physics.optics2016

High Performance Atomically Thin Flat Lenses

Han Lin, Zai-Quan Xu, Chengwei Qiu +2

We experimentally demonstrate ultrathin flat lenses with a thickness of 7 Ã , which corresponds to the fundamental physical limit of the thickness of the material, is fabricated in…

cs.AI2026

RUMAD: Reinforcement-Unifying Multi-Agent Debate

Chao Wang, Han Lin, Huaze Tang +2

Multi-agent debate (MAD) systems leverage collective intelligence to enhance reasoning capabilities, yet existing approaches struggle to simultaneously optimize accuracy, consensus…

physics.optics2018

A nanophotonic all-optical diode for non-reciprocal transmission of circularly polarized lights

Hongming Fei, Min Wu, Han Lin +4

All optical diodes (AODs) play an important role in quantum optics and information processing, in which the information is encoded by photons. Only circularly polarized lights are…

astro-ph.HE2021

SN 2018hfm : A Low-Energy Type II Supernova with Prominent Signatures of Circumstellar Interaction and Dust Formation

Xinghan Zhang, Xiaofeng Wang, Hanna Sai +13

We present multiband optical photometric and spectroscopic observations of an unusual Type II supernova, SN 2018hfm, which exploded in the nearby (d = 34.67 Mpc) dwarf galaxy PGC 1…