2024年5月26日, 09:15-09:40 | 中国 · 杭州
王赫 (He Wang)
20分钟 大概20页
我在北师大的求学和渊源(1)
引力波天文学与AI
人才培养
AlphaGo
围棋机器人
AlphaTensor
发现矩阵算法
AlphaFold
蛋白质结构预测
验证数学猜想
Gravitational waves generated by binary black holes system
GW detector
引力波探测打开了探索宇宙的新窗口
不同波源,频率跨越 20 个数量级,不同探测器
多信使天文学
DOI:10.1063/1.1629411
©Floor Broekgaarden (repo)
The first GW event of GW150914
引力波观测数据
噪声: 非高斯 + 非稳态
(地面引力波探测) 信噪比极低,通常约为噪声幅度的1/100(-60分贝)
(空间引力波探测) 在任务观测期间接收到的所有引力波信号的叠加(例如:\(10^4\) 个双星黑洞系统,\(10\sim10^2\) 个超大质量黑洞,以及\(10\sim10^3\) 个极端质量比旋近系统等)。
匹配滤波方法
高斯且稳态噪声环境下,提取弱信号的最优算法
假设:引力波观测数据 = 高斯稳态噪声 + 某引力波信号
\(d(t) = n(t) + h(t)\)
LIGO-VIRGO-KAGRA
LISA / Taiji project
2016年,AlphaGo 第一版发表在了 Nature 杂志上
2021年,AI预测蛋白质结构登上 Science、Nature 年度技术突破,潜力无穷
2022年,DeepMind团队通过游戏训练AI发现矩阵乘法算法问题
《达摩院2022十大科技趋势》将 AI for Science 列为重要趋势
“人工智能成为科学家的新生产工具,催生科研新范式”
2023年,DeepMind发布AI工具GNoME (Nature),成功预测220万种晶体结构
AI for Science:为科学带来了模型与数据双驱动的新的研究范式
AI + 数学、AI + 化学、AI + 医药、AI + 量子、AI + 物理、AI + 天文 ...
AlphaGo 围棋机器人
AlphaTensor 发现矩阵算法
AlphaFold 蛋白质结构预测
验证数学猜想
Pioneering works utilizing CNN
AI for Science \(\rightarrow\) AI for GW Astronomy
Exported: Oct, 2023 (in preparation)
PRL, 2018, 120(14): 141103.
PRD, 2018, 97(4): 044039.
Matched-filtering Convolutional Neural Network (MFCNN)
MLGWSC-1
The majority of AI algorithms used for testing are highly sensitive to non-Gaussian real noise backgrounds, resulting in high false positive rates.
(MFCNN group) H.W., et al. PRD (2023)
CL.M., W.W., H.W., et al. PRD (2022)
Ensemble learning
Leverages statistical approaches to utilize more information for making informed decisions by combining multiple models.
Real-time GW searches for GW150914
H.W., et al. PRD (2020)
Expanding the dimension of the output
CL.M., W.W., H.W., et al. PRD (2023)
Beyond Speed: Generalization and Discovery in GW Detection
He Wang, et al. MLST. 5, 1 (2024): 015046.
Challenges in Model Interpretability
He Wang, et al. MLST. 5, 1 (2024): 015046.
GW151226
GW151012
LVK. arXiv:1602.03839
Credit: LIGO Magazine.
Traditional parameter estimation (PE) techniques rely on Bayesian analysis methods (posteriors + evidence)
Bayesian statistics
Data quality improvement
Credit: Marco Cavaglià
LIGO-Virgo data processing
GW searches
Astrophsical interpretation of GW sources
PRL 127, 24 (2021) 241103.
PRL 130, 17 (2023) 171403.
Nature Physics 18, 1 (2022) 112–17
HW, et al. Big Data Mining and Analytics 5, 1 (2021) 53–63.
A diagram of prior sampling between feature space and physical parameter space
Simulation-Based Inference (SBI)
PRL 127, 24 (2021) 241103.
PRL 130, 17 (2023) 171403.
Real-time gravitational wave science with neural posterior estimation
Sampling with prior knowledge for high-dimensional gravitational wave data analysis
He Wang, et al. Big Data Min. Anal. (2021)
PRD 108, 4 (2023): 044029.
Neural Posterior Estimation with Guaranteed Exact Coverage: The Ringdown of GW150914
arXiv:2310.13405, LIGO-P2300306
Cosmological Inference using Gravitational Waves and Normalising Flows
Fast Parameter Inference on Pulsar Timing Arrays with Normalizing Flows
arXiv:2310.12209
Normalizing Flows as an Avenue to Studying Overlapping Gravitational Wave Signals
PRL 131, 17 (2023): 171403.
Angular Power Spectrum of Gravitational-Wave Transient Sources as a Probe of the Large-Scale Structure
M. Du, B. Liang, HW*, P. Xu, Z. Luo, Y. Wu*. SCPMA 67, 230412 (2024).
Multimodality in extrinsic parameters
Computational performance
10000 samples in 2.7 sec
Rapid PE for Space-borne GW Detection
数字信号处理
R.C. Cofer, Benjamin F. Harding, in Rapid System Prototyping with FPGAs, 2006
Dieter Rasch, Dieter Schott. Mathematical Statistics, (2018)
数理统计
理论基础:
引力波物理
数字信号处理
数理统计
编程基础:
硬件基础:
Miller, M.C., Yunes, N. The new frontier of gravitational waves. Nature 568, 469–476 (2019).
引力波物理与引力波天文学
理论基础:
引力波物理 (pycbc, lalsuite, lisacode, bilby, ... )
数字信号处理 (scipy, stat, ...)
数理统计 (bilby, emcee, ptemcee, ptmcmc, …)
编程基础:
Python (numpy, pandas; matplotlib; ...),
AI (scikit-learn, XGBoost, PyTorch, TensorFlow, JAX, ...)
Linux (docker, github, bash, vim, emacs …)
硬件基础:
主板、内存,GPUs,显存 ...
2. 引力波太极实验室:《引力波数据探索:编程与分析实战训练营》 2023.11 - 2024.1
2. 引力波太极实验室:《引力波数据探索:编程与分析实战训练营》 2023.11 - 2024.1
2. 引力波太极实验室:《引力波数据探索:编程与分析实战训练营》 2023.11 - 2024.1
AI or Bayes
Text-to-image
"A running dog"
AI or Bayes
Text-to-image
"A corgi running on the street"
A picture is worth a thousand words.
A fraction of a thousand words.
Credit: 李宏毅
"A running dog"
2310.12528
Pipeline | Targets | Programing Language (sampling method) | Comments |
---|---|---|---|
GLASS (Littenberg&Cornish 2023) |
Noise, UCB, VGB, MBHB |
C / Python (TPMCMC / RJMCMC) | noise_mcmc+gb_mcmc+vb_mcmc+global_fit |
Eryn | UCB | Python (TPMCMC / RJMCMC) | Mini code for UCB case |
PyCBC-INFERENCE | MBHB | Python (?) | Unavailable |
Bilby in Space / tBilby | MBHB / ? | ? / Python? (RJMCMC) | Unavailable |
Strub et al. | UCB | ? (GP) | Unavailable / GPU-based |
Zhang et al. (LZU) | UCB | ? (PSO) | MLP |
Balrog | MBHB | ? |
(Sec.8.6 Red Book)
Global Fit
Technical challenges:
Neural density estimation
Ref:
Neural density estimation
nflow