- 博客(33)
- 收藏
- 关注
原创 DDPM( Denoising Diffusion Probabilistic Model )
介绍 DDPM( Denoising Diffusion Probabilistic Model )的算法原理
2022-07-20 16:01:31
3779
5
原创 JTVAE( Junction Tree Variational Autoencoder )
Junction Tree Variational Autoencoder for Molecular Graph GenerationYear: 2018Authors: Wengong Jin, Regina Barzilay, Tommi JaakkolaJournal Name: ICMLInnovation
2022-04-10 10:00:55
2202
原创 VGAE( Variational Graph Auto-Encoder )
Variational Graph Auto-EncodersYear: 2016Authors: Thomas N. Kipf, Max WellingJournal Name: NIPSDefinitions定义含有 N=∣V∣N=|V|N=∣V∣ 个节点的无向图 G=(V,E)\mathcal{G}=(\mathcal{V}, \mathcal{E})G=(V,E) ,邻接矩阵 AAA 和度矩阵 DDD ,再引入 N×FN \times FN×F 大小的隐变量矩阵 Z=[z1;z2;...;
2022-03-18 10:54:19
1047
1
原创 De Novo Prediction of RNA 3D Structures with Deep Learning
De Novo Prediction of RNA 3D Structures with Deep LearningYear: 2022Authors: Julius Ramakers, Christopher Frederik Blum, Sabrina K¨onig, Stefan Harmeling, Markus KollmannJournal Name: bioxiv1 Innovation结合自回归深度生成模型、蒙特卡罗树搜索和分数模型预测 RNA 三维折叠结构。2 Method
2022-03-12 11:35:05
2037
原创 PINN potentials
Physically informed artificial neural networks for atomistic modeling of materialsYear: 2019Authors: G. P. Purja Pun, R. Batra, R. Ramprasad & Y. MishinJournal Name: Nature CommunicationsInnovation扩展物理模型,使其具有广泛适用性,即在训练集覆盖不到的数据中也能适用。引入与局部结构参数 Gil
2022-02-17 13:31:45
497
原创 RNAcmap--predicting contact maps of RNAs by evolutionary coupling analysis
RNAcmap: a fully automatic pipeline for predicting contact maps of RNAs by evolutionary coupling analysisYear: 2021Authors: Tongchuan Zhang, Jaswinder Singh, Thomas Litfin, Jian Zhan, Kuldip Paliwal and Yaoqi ZhouJournal Name: BioinformaticsMotivation
2022-01-26 15:42:29
477
原创 ARES( Atomic Rotationally Equivariant Scorer )
Geometric deep learning of RNA structureYear: 2021Authors: Raphael J. L. Townshend, Stephan Eismann, Andrew M. Watkins, Ramya Rangan, Maria Karelina, Rhiju Das, Ron O. DrorJournal Name: ScienceDatasetBackgroundMethodEquivariant convolution等变卷积基于滤波
2022-01-24 16:13:32
557
原创 GFlowNet Foundation 笔记(五)
系列文章GFlowNet Foundation 笔记(一)GFlowNet Foundation 笔记(二)GFlowNet Foundation 笔记(三)GFlowNet Foundation 笔记(四)期望的奖励和奖励最大策略Def 37. 对于在终止状态上的任意分布 Pπ(s)P_{\pi}(s)Pπ(s) ,期望奖励( expected reward ) 为VPπ(s)=EPπ(S)[R(S)∣S≥s]=∑s′≥sR(s′)Pπ(s′∣s≤s′)V_{P_{\pi}}(s) =
2022-01-07 20:32:33
810
原创 GFlowNet Foundation 笔记(四)
系列文章GFlowNet Foundation 笔记(一)GFlowNet Foundation 笔记(二)GFlowNet Foundation 笔记(三)确定环境和随机环境中的策略Def 34. 策略( policy ) π:A×S↦R\pi: \mathcal{A} \times \mathcal{S} \mapsto \Rπ:A×S↦R 为概率分布 π(a∣s)\pi (a | s)π(a∣s) 。其中,行动 a∈Aa \in \mathcal{A}a∈A ,定义 A(s)\mathcal
2022-01-07 09:35:06
1228
原创 GFlowNet Foundation 笔记(三)
系列文章GFlowNet Foundation 笔记(一)GFlowNet Foundation 笔记(二)条件流与自由能Def 24. 已知自由能 F(s)\mathcal{F}(s)F(s)e−F(s)=∑s′:s′≥sR(s′)=∑s′:s′≥se−F(s′)e^{-\mathcal{F}(s)} = \sum_{s': s' \ge s} R(s') = \sum_{s': s' \ge s} e^{-\mathcal{F}(s')}e−F(s)=s′:s′≥s∑R(s′)=s′:s
2022-01-04 10:51:55
733
原创 GFlowNet Foundation 笔记(二)
学习流量Def 18. GFlowNet 可用 (F^(s),P^F(st+1∣st))(\hat{F}(s), \hat{P}_F(s_{t+1} | s_t))(F^(s),P^F(st+1∣st)) 表示。从终止流量估计转移概率终止流量对应终止奖励函数 RRRR(s)=F(s→sf)R(s) = F(s \rightarrow s_f)R(s)=F(s→sf)推论3. 上面的式子可以推导出总流量Z=F(s0)=F(sf)=∑s∈Par(sf)R(s)Z = F(s_0) =
2022-01-01 15:41:32
707
1
原创 GFlowNet Foundation 笔记(一)
马尔可夫流量的测量轨迹( trajectories )和流量( flows )Def 1. 一个有向图用 (S,A)(\mathcal{S}, \mathbb{A})(S,A) 表示,其中 S\mathcal{S}S 为状态的集合, A\mathbb{A}A 为 S×S\mathcal{S} \times \mathcal{S}S×S 大小的有向边的子集。 A\mathbb{A}A 中的元素表示为 s→s′s \rightarrow s's→s′ ,叫做 边缘( edges ) 或 转移( transi
2021-12-28 11:49:37
1339
原创 坐标上升变分推断( Coordinate Ascent Variational Inference, CAVI)
变分推断是为了近似获得 P(Z∣X)P(Z | X)P(Z∣X) ,即隐状态的后验分布。logP(X)=logP(X,Z)−logP(Z∣X)=logP(X,Z)q(Z)−logP(Z∣X)q(Z)\begin{aligned} log P(X) &= log P(X, Z) - log P(Z | X) \\ &= log \frac{P(X, Z)}{q(Z)} - log \frac{P(Z | X)}{q(Z)}\end{aligned}logP(X)=logP(X,Z
2021-12-19 22:46:51
806
原创 估计参数的均方误差
花书 5.4.45.4.45.4.4 中说均方误差( MSE )度量着估计 θ^\hat{\theta}θ^ 和真实参数 θ\thetaθ 之间平方误差的总体期望偏差,但没有进行证明,现将推导过程展示在下面MSE=E[(θ^−θ)2]=E[((θ^−E(θ^))+(E(θ^)−θ))2]=E[(θ^−E(θ^))2]+E[(E(θ^)−θ)2]+2E[(θ^−E(θ^))(E(θ^)−θ)]=E[(θ^−E(θ^))2]+(E(θ^)−θ)2+2(E(θ^)−E(θ^))(E(θ^)−θ)=Var(θ^)
2021-11-30 15:08:12
1698
原创 方向导数 directional derivative
最近在看花书,其中 4.3 中提到了方向导数当 α=0\alpha = 0α=0 时, ∂∂αf(x+αu)=uT∇xf(x)\frac{\partial}{\partial \alpha}f(\textbf{x} + \alpha \textbf{u}) = \textbf{u}^T \nabla_xf(\textbf{x})∂α∂f(x+αu)=uT∇xf(x) 是怎么得出的?根据全微分∂f(x+αu)=∂f(x1+αu1)∂x1⋅∂(x1+αu1)+⋯+∂f(xn+αun)∂xn⋅∂(xn+
2021-11-27 19:12:34
832
1
原创 传统图生成方法
系列文章《Graph Representation Learning》笔记 Chapter2《Graph Representation Learning》笔记 Chapter3《Graph Representation Learning》笔记 Chapter4《Graph Representation Learning》笔记 Chapter5《Graph Representation Learning》笔记 Chapter6传统方法概述我们可以将生成过程指定为计算 P(A[u,v]=1)P(A
2021-11-25 21:11:50
998
原创 《Graph Representation Learning》笔记 Chapter6
系列文章《Graph Representation Learning》笔记 Chapter2《Graph Representation Learning》笔记 Chapter3《Graph Representation Learning》笔记 Chapter4《Graph Representation Learning》笔记 Chapter5目录Applications and Loss FunctionsGNNs for Node ClassificationGNNs for Graph Cla
2021-11-25 17:08:03
580
原创 《Graph Representation Learning》笔记 Chapter5
系列文章《Graph Representation Learning》笔记 Chapter2《Graph Representation Learning》笔记 Chapter3《Graph Representation Learning》笔记 Chapter4目录Permutation invariance and equivarianceNeural Message PassingOverview of the Message Passing FrameworkPermutation invar
2021-11-25 15:43:43
2643
原创 BLAST
Basic Local Alignment Search ToolYear: 1990Authors: Stephen F. Altschul, Warren Gish, Webb Miller, Eugene W. Myers and David J. LipmanJournal Name: Journal of Molecular BiologyAbstract一个快速进行序列比较的新方法:基本局部比对搜索工具( basic local alignment search tool, BLAST
2021-11-24 19:40:05
303
原创 LinearPartition
LinearPartition: linear-time approximation of RNA folding partition function and base-pairing probabilitiesYear: 2020Authors: He Zhang, Liang Zhang, David H. Mathews and Liang HuangJournal Name: BioinformaticsMotivation传统分割方法的复杂度与序列长度呈三次方关系。Research
2021-11-22 21:39:35
711
原创 2D-LSTM
LSTMUnderstanding LSTM Networks 和 人人都能看懂的LSTM 这两篇文章介绍了 LSTM 的原理。2D-LSTM2D-LSTM 是作用于三维输入( W×H×DW \times H \times DW×H×D )的 LSTM ,分别取横向和纵向上一时刻的隐藏状态和输出作为该时刻的输入,如下图所示数据传播的顺序依靠对角线原则,如下图所示图中的数字表示计算的顺序。下图展示了 2D-LSTM 单元的结构,蓝线表示与标准单元不同的地方。上图中 xj,ix_{j, i}
2021-11-17 11:24:51
4898
原创 torch.nn.LSTM
什么是 LSTMUnderstanding LSTM Networks 和 人人都能看懂的LSTM 这两篇文章介绍了 LSTM 的原理。本文的着重点在于 LSTM 的输入输出维度以及 torch.nn.LSTM 的使用。LSTM 的输入输出首先来看这张图input 序列的长度为 LLL ,包括 x1,x2,...,xn−1,xn(n=L)x_1, x_2, ..., x_{n-1}, x_n(n=L)x1,x2,...,xn−1,xn(n=L) ,每个输入 xnx_nxn 的维度数 in
2021-11-15 22:30:05
2370
原创 向量空间 vector space
向量空间表示为 R1,R2,R3,R4,...,Rn\bf{R}^1, \bf{R}^2, \bf{R}^3, \bf{R}^4, ..., \bf{R}^nR1,R2,R3,R4,...,Rn 。 Rn\bf{R}^nRn 表示 nnn 维向量集合所组成的空间,称为二维向量空间。但是,并不是所有集合组成的空间都能称作向量空间,必须满足以下 2 个条件集合中的两个向量相加,所得的新向量依然在这个集合内。集合中的任一向量与一标量相乘,所得的新向量依然在这个集合内。举两个例子( from mit公开课
2021-11-13 23:11:38
1695
原创 SPOT-RNA
RNA secondary structure prediction using stochastic context-free grammars and evolutionary historyYear: 1999Authors: B. Knudsen and J. HeinJournal Name: BioinformaticsMotivation最近,一些进化模型被应用于结构预测中,这些模型忽略了一些相关信息。什么叫进化模型?什么样的相关信息?Research Objective结合 s
2021-11-10 16:51:34
1028
原创 SCFGs
Stochastic context-free grammars for tRNA modelingYear: 1994Authors: Yasubumi Sakakibara, Michael Brown, Richard Hughey, I.Saira Mian, Kimmen Sjolander, Rebecca C.Underwood and David HausslerJournal Name: Nucleic Acids ResearchResearch Objective通过类似于构
2021-11-09 13:52:46
472
原创 MXfold2
RNA secondary structure prediction using deep learning with thermodynamic integrationYear: 2021Authors: Kengo Sato, Manato Akiyama & Yasubumi SakakibaraJournal Name: Nature CommunicationsMotivation在多参数模型种经常会出现过拟合现象Research Objective实现一种更具鲁棒性的结构预
2021-11-07 21:16:50
479
原创 Loss-augmented Max-margin Constraint Generation(LAM-CG) algorithm
Computational approaches for RNA energy parameter estimationYear: 2010Authors: MIRELA ANDRONESCU, ANNE CONDON, HOLGER H. HOOS, DAVID H. MATHEWS, and KEVIN P. MURPHYJournal Name: BIOINFORMATICSMotivation将最大间隔应用于CG模型中Research Objective改进估计能量参数的优化方法。B
2021-11-06 20:51:21
153
原创 Dynamic programming algorithm
Optimal computer folding of large RNA sequences using thermodynamics and auxiliary informationYear: 1981Authors: Michael Zuker and Patrick StieglerJournal Name: Nucleic Acids ResearchMotivation将动态规划算法与热动力学数据结合。Research Objective基于动态规划算法提出一种更有效、快速、并且
2021-11-06 14:37:09
371
原创 Constraint generation(CG) approach
Efficient parameter estimation for RNA secondary structure predictionYear: 2007Authors: Mirela Andronescu, Anne Condon, Holger H. Hoos, David H. Mathews, and Kevin P. MurphyJournal Name: BioinformaticsMotivation基于自由能的RNA二级序列预测模型Turner99,没有有效、快速的参数估计方
2021-11-05 16:15:27
197
原创 《Graph Representation Learning》笔记 Chapter4
系列文章《Graph Representation Learning》笔记 Chapter2《Graph Representation Learning》笔记 Chapter3目录Reconstructing muti-relational dataLoss functionCross-entropy with negative samplingMax-margin lossMulti-relational decoderReanslational decodersMulti-linear dot p
2021-10-16 16:11:30
432
原创 《Graph Representation Learning》笔记 Chapter3
系列文章《Graph Representation Learning》笔记 Chapter2目录An Encoder-Decoder Perspective编码 The Encoder解码 The DecoderOptimizing an Encoder-Decoder ModelOverview of the Encoder-Decoder ApproachFactorization-based approachesLaplacian eigenmaps内积方法 Inner-product metho
2021-09-29 11:46:01
363
原创 《Graph Representation Learning》笔记 Chapter2
目录节点特征node degreenode centrality集聚系数 The clustering coefficient图特征Weisfieler-Lehman kernel邻域重叠检测 Neighborhood Overlap Detection局部重叠度量 Local overlap measure全局重叠度量 Global overlap measureKate indexLeicht, Holme, and Newman(LHN) similarity随机游走法 Random walk met
2021-09-21 15:27:33
628
空空如也
空空如也
TA创建的收藏夹 TA关注的收藏夹
TA关注的人