Archives

DeepSeek has great influence during the past two months. It is open-weight (not open source code) with fascinating algorithm and hardware level optimization. It is both effective and efficient. I am curios in its technical innovation and here is a slide presented in our group meeting

> 云想衣裳花想容， 春风拂槛露华浓。 若非群玉山头见， 会向瑶台月下逢。

孟孟24岁生日的时候我整理了一本小书，用拾柒制作的，印刷得挺不错，书名取自高中的时候写给孟孟的表白的文章，叫早安晨曦

OUR PAPER WAS PUBLISHED on ISBI 2020 as Best Paper Finalist LINK


想写这个很久了，题目就想叫做几件小事，因为和孟孟在一起这么多年，发生了很多我们心中很珍重的大事，还有离别半年来一件件在我记忆里反而愈发清晰和留恋的无数的小事，我无法一一列举，只能在空中胡乱一抓，把几件自然就涌入脑子中的小事记下来。这篇文章的提纲和一些细节我一个月前就列下来了，一直想找一个最好的时机和心情写下来，无奈平时太忙，总是被各种事情催着，今天晚上又翻了翻以前随手拍的照片和无数的聊天记录，决定不再拖延了，把文章完成，来提醒自己，这样的美好的日子，在自己的生命里，发生过无数次，不能忘记。

Here I summarized a very exciting differentiable digital signal processing tool proposed by Google's team in Jan 2020. The tool is especially useful in audio waveform reconstruction and some other related task.

接下来是孟孟写的迪士尼一日游记~

打车等了一个多小时，跨年夜真的恐怖如斯，最近几天过的十分平和喜乐，回来四五天了，一直窝在短时的住处和孟孟的学校，享受着半年多以来最舒适惬意的一段日子。想想我最爱的状态其实就是安静简单+和最爱的人每天见面，这几天满足了我的一切心愿，用这样的方式迎接新年是我所爱的。

昨天看到同学在群里分享了一个很好玩的游戏，叫PhD 2048，我就在上面做了些DIY改了一个mengmeng 2048出来

现在真的很少写长一点的文字了，初中在语文老师的督促下每周都有摘抄赏析周记，高中也知道写周记写情书，大学之后就只会在朋友圈指点江山，发些零散的感慨了，但是在这种时刻又很容易被感染，忍不住想写点说点什么，来纪念正值壮年的祖国七十周岁生日快乐。

First two weeks in new semester and video lab

最后一站：Tokyo in Japan

第四-六站：Osaka, Nara and Kyoto in Japan

第三站：Phuket in Thailand

第二站：Chiang Mai in Thailand

第一站：Bangkok in Thailand

First two weeks in NYC and BK

2019/6/18 是我的女朋友孟孟的毕业典礼，一大早我们就准备好要拍美美的照片了，因为觉得（~~懒~~）孟孟太美了，所以图片都没有P就放上来了，我一直觉得没有P的图片最自然最好看~~省事~~，非常符合这种庄重又美丽的时刻。

我是想好好写写我的本科生活的回忆的，可是一想到要写好长好长就感觉好累，希望这是个长长的故事，我会在未来回来继续它。

> The original six avengers, each carrying a stone and fight with Thanos. Thanos snaps his fingers first and each avenger has a 50% probability to die. If all of the avengers die, thanos wins. If not, one of the alive avengers snap fingers and thanos has 50% percent to die. The fights will continue

> edit (2019-6-13): the final presentation is over, the last step to graduate is over!

“整理这些碎碎念让我更好的感觉到我还活着”——2019.4.23

Who don't like android?

A cute raspberry is a fantastic toy to play with.

The first assignment of Digital Image Processing course (2019 spring in Tsinghua)

可以说我的申请季过得相当充实而有意义，在这个忙乱的申请季，通过种种的思索和抉择，让我慢慢找寻到了未来的方向。

It is a small tool converting LaTeX style code to MathML which can be used in places like word.

I have been responsible for several (part of the ) books.

Parts of this works difficulty is its steps and precision it requires. Since our exRNA-seq is different (it has less RNA reads), we have to explore the mapping sequence. We determine sequential mapping order, using peak calling because most long RNAs are fragments in blood. We also did matrix proces

We then turn to predict RNA motif (especially RNA-protein interaction related motif). It is another harder problem compared with "simply" applying deep neural network to 1D and 2D structure data. We develop the mixture model for PWM optimization. We explore the possibility to use VAE for unsupervise

For Synapse project, I work with Zudi and Donglai, we now rank No.1 in the CREMI contest. And we will submit a paper to CVPR this November.

We proposed a new mixed linear model Mixed Ridge with Fast cross validation to pick parameters. We also propose Metric Regressor for few shot learning.

We collaborate with Jun Li in New York University.

The related jupyter file

The related jupyter file

The related jupyter file

The related jupyter file

They are parts of the contribution during our return to high school program. As the team leader, I organized a team with more than three hundreds elite students from top universities in China to return to high school. We built a platform to publish life and thoughts about high school life in a daily

I have spent many time written several chapters of the books. For Basic teaching book I contributed more than 1/3 of the book.

I mainly focus on semantic segmentation and potentially related papers.

As a student from life science background with machine learning and deep learning skills, I have a broader mind to think about medical image problems from different views. From my perspectives, deep learning aided medical image analysis will soon have a wide range of application.

- SummerIntern/synapticpartner at master · james20141606/SummerIntern · GitHub - Synapse polarity - SummerIntern/synapsecluster at master · james20141606/SummerIntern · GitHub

- SummerIntern/synapticpartner at master · james20141606/SummerIntern · GitHub - Synapse polarity

Also I finished another synapse prediction and synaptic polarity prediction work during summer intern in Hanspiter Lab

When I prepared my presentation about Wittgenstein, especially chapter RELUCTANT PROFESSOR, I write down some of my thoughts towards Wittgenstein, some of them is biased, because I read something which reveals the weakness and defect about him:

我记下了一些一下子就能想到的事情，还有太多太多细节都回荡在脑海里，如果有时间的话，应该仔细地写一写这些宝贵的回忆。

Artificial intelligence means using program to accomplish tasks human can do by simulating human’s brain. The purpose of artificial intelligence is to bypass the brain and body to achieve a full understanding of human intelligence. The idea can be traced back to Turing machine (we can already see so

The thirteenth assignment of Linear Regression. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

本文写于2016年末，作为《脑科学与人工智能》一课的作业

The fifth assignment of Multivariate Statistics. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

The fifth assignment of Multivariate Statistics. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

The fifth assignment of Multivariate Statistics. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

The fifth assignment of Multivariate Statistics. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

The fifth assignment of Multivariate Statistics. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

显然既然automator还可以用shell脚本，能低成本实现的东西还是很多的。

But I will write it down. I am strong when facing anything else, I can always find ways to solve the hard problems, I find my biggest fear happens to be something I should try to solve. I may find some way to hide my fear, to do something else, to earn many money, to live a happy, ignorant life. But

The nineth assignment of Linear Regression. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

The eleventh assignment of Linear Regression. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

The twelveth assignment of Linear Regression. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

It contains my assignment in Bioinformatics course talking about Differential Expression and this semesters' code

去年最火的微信机器人似乎是GitHub - littlecodersh/ItChat: A complete and graceful API for Wechat. 微信个人号接口、微信机器人及命令行微信，三十行即可自定义个人号机器人。，尤其是这种文章用微信控制深度学习训练：中国特色的keras插件还是相当有趣的。今天又看到一个功能很全的微信接口https://github.com/youfou/wxpy

The fourth assignment of Multivariate Statistics. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

看着文章的第一反应是如果生孩子这么辛苦和危险，要想办法找到更好的条件减轻风险，第二反应是妈妈真不容易，很勇敢，很了不起，勇敢地把我生了下来。我想世界的话语权长期被男性把持着，所以一直低估了女性在生育过程中经历的漫长的、折磨人的却又自愿的痛苦。如果细细回想，细细品味，从男人的角度看那些令人揪心的描述，想象漫长的数十个月甚至更久的煎熬和痛苦，实在是值得每个男人潸然泪下，为自己的母亲、自己的妻子和自己的孩子。

The first version has been finished, I am really happy to hear that grandpa is excited to see what I have done. He is motivated to write more about his life. I hope there will be more inspiring and touching old stories soon.

snakemake是一个用来编写任务流程的工具，用python写的，因此其执行的流程脚本也比较通俗易懂，易于理解，可以看做用户友好版的make。

今天探索了一下如何添加个人网站到百度和google搜索引擎中

在排版的文章里讲过论文与排版 | James Chen’s Blogs

For example, I even set one alias: srczsh='source ~/.zshrc' to enable source more convienient.

Conda是个好东西，以前的文章介绍过Setup and Linux | James Chen’s Blogs，因为好几台机器，因为某些库，以及python2和3的问题，重装倒腾过N次，于是倒腾出来一个标准流程，记录如下。

下面就讲一下jupyter的一些高级玩法。

在这里分享一下各种相关资源和知识

之前用过一些，比如： - 帮老板写过书里的一章(5.1 eMaize Challenge · 生物信息学实践) - training课的一些notes(Introduction · Bioinfo Training - Shared - 2018)，后来又移到github的wiki上了(Home · lulab/training Wiki · GitHub) - 还把爷爷的回忆录也用markdown整理了一遍放到gitbook上(爷爷回忆录)，（一文多用，之前也已经放到blog上面了），排版比word辛辛苦苦排来排去随便一弄就全乱了是强太多了。

The eigth assignment of Linear Regression. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

Some notes and thoughts concerning Ludwig Wittgenstein: The Duty of Genius by Monk, Ray. To be honest, I first found Wittgenstein very mysterious and great (part of the reason is his wonderful german name). But now I feel little disappointed. I see more a crazy and strange man rather than a genius.

ALL the assignments of Statistic Inferences. The assignment is written in Rmarkdown/LaTeX, LaTeX is a classic formula editor but a little hard to use intensely. Rmarkdown is introduced by TA and it is a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes runni

The seventh assignment of Linear Regression. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

The sixth assignment of Linear Regression. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

The fourth assignment of Linear Regression. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

The third assignment of Multivariate Statistics. The assignment is written in Rmarkdown, a smart syntax supported by RStudio helping with formula, plot visualization and plugin codes running.

就像pdf文档里所说的那样，这组诗作于2013年春⽇在郑州外国语学校就读期间，我十七岁，孟孟年方二八，正是豆蔻年华，如今看起来有些羞涩的文字却是那个时候的真情流露，并不觉得夸张。人们总是喜欢用青涩形容自己年轻的岁月，我却一直秉持着大胆地做自己想做的事情的原则，因此青春中做了很多美好的值得回味的，不落于俗套的美好事情。这一路来所做所想，都是自己所爱，能够在自己多年来心仪的学校继续和多年来心爱的姑娘经历生活的美好与平淡，实在是人生的幸运。

期中作业之一是写一篇邓小平时代的读后感，说实话这种书是实在没空读了，虽然粗略地翻了几章，十分吸引人，但是这两年真的越来越讨厌写文科式的论文，瞎胡诌凑字数曾经也是我作为理科生的优势，但是这一两年对这种风格的文章：东拼西凑，无病呻吟，迷茫又自负的写作非常地厌恶。因为就想玩点花样，做点简单的文本数据挖掘凑凑字数，虽然多花了很多时间，但是毕竟很有意思，有意思的事情就不算浪费时间对吧，没有意思的事情，哪怕一分钟也是对生命的浪费呢。

分享一点setup和linux的东西，包括使用git，使用支持markdown的笔记软件Bear，Anaconda的一些使用技巧以及jupyter在服务器上的设置。也可以在这里找到更多分享

由于jupyter notebook的强大的展示功能，本教程还用jupyter notebook组织且运行，可以获得更好的学习效果，代码在这里,欢迎取用。在这里我简单介绍了如何配置jupyter，在Deep Learning tutorial中我也强烈推荐了jupyter，并且介绍了很多基于jupyter的资源，强烈建议尝试一下。

本文章同时被整理成一本小书，放置在我的gitbook账户下，点击这里可以阅读。

我出生在石桥街西夹后布袋街外祖母家那个巷子里，是一九四零年(民国二十九年)生，赶上三十年年成那年。我出生后家里有四口人，大哥已经两岁。在集镇上住，家里没地没房，不做生意，生存十分困难。后来经人介绍，父亲用卖菜的筐一头一个孩子，挑着我们去白河东沙山给地主彭山种地。地主给了草房两间，几亩薄地，生活勉强过得去。日本侵华后战乱频起，又逢灾年（指1942年七月到1943年春天的那场大灾荒，河南受灾总人数达1200万人，约三百万人死亡），祖母不愿骨肉分离，我们一家四口只好又两手空空搬到祖母借住地薛庄去（魏庄西边西边的那个庄）。

入学后我被分到了56丁班，和我后来的妻子是同班同学，她从小学就学习刻苦，成绩优秀，是初中部的学习部长，那时我们是纯洁的同学关系，很少有接触，几乎没有怎么说过话。

父亲病倒后水库指挥部通知家人，去云阳水库接回父亲。家里只有我可以担当此仁，此时我已经十八岁，刚刚成年。接到通知的时候我害怕极了，父亲可千万不能出事啊。我恨不得自己能赶快长出一双翅膀飞过去接他回家。

我在一九五九年夏季初中毕业，学校要求毕业生要主要报考高中、师范。高中是绝对上不起的，当教师是臭老九，经过反复的斟酌，我选择了当时自己觉得最受人崇拜的专业：农机化专业，这可是毛主席提出的四化之一，就这样我顺利录取到了郑州农机化专科学校（1960年，郑州农业机械化专科学校并入河南农学院（现河南农业大学），组建河南农学院农业机械化分院），我仿佛已经能够想象出自己开着铁牛奔驰在祖国广阔原野上的场景了。

回乡后的我很快就适应了农村的生活，我是地道的农民的儿子，但是我心里还是惦记着上学时的生活，惦记着我的老同学，他们现在生活的怎么样，他/她们在干什么呢？老家尹店每年农历四月初八的庙会规模大，总是有两台戏，约老同学来赶庙会、叙叙旧应该是不错的选择！于是同班同学王若平和赵玉珍应约前往，逛完庙会又去了我的家里，父母热情地接待了她们，也让她们感动不已。

一九六二年的农历闰四月，家里连一点可以吃的东西都没有了。已经怀了身孕的妻子只能靠着白水煮红薯干充饥，想吃点别的什么食物都是奢望。老家每年四月八号的庙会总会有亲朋好友和同学们来家里，断炊了还拿什么招待呢？我决定跟随邻居，去燕尾山拾柴变卖来赚些钱买点粮食。

我和若平两人背景相同，志同道合，结婚后朝夕相处，可以永不分离，两个人形影不离，生活觉得甜蜜极了，结伴扛着农具干活去，在乡下的生活倒也过得有滋有味。我的长子陈大军出生后，全家都高兴坏了。父亲说，长子得了长孙，家里人丁兴旺了起来，再困难也得办几桌酒席，让亲朋好友齐聚一堂，共享快乐。父亲是非常能干厉害的人，在那么困难的条件下（三年灾害刚刚过去），操办了全席酒宴，我由衷地感激父亲。

一九六四年八月中旬，我和妻子顺利毕业了，回宛后住在现在的新华宾馆等地方人事部门分配工作，妻子被分配到原安皋拖拉机站，我被安排到茶庵拖拉机站，过着两地分居的生活。 我在茶庵期间，单位的一把手杨永江是原来的茶庵区副区长，人很能干精明。他是姑婆家（槐树湾姑爷妹妹的老公）亲戚，待我非常的好，我就和他住在一个房间，他既是领导，也是朋友、同事，我们无话不说，因此工作上我就不怕了，谁也不敢欺负我。一九六四年八月十五日到达工作岗位后，我赶上了发工资，领到了人生的第一笔工资，成了国家的正式干部。

那是上世纪第二个十年发生的故事，年代久远，记忆都已褪去了鲜活的颜色，就像黑白照片一样，边缘还笼罩着一层薄雾。故事是老家的一位年迈的智者在弥留之际讲述的，这个故事在他心中藏了很多年，这个充满了残忍的谋害、热血的沸腾、正义的伸张的故事，糅合在那个特殊的年代，被人藏在心里，藏得太久，当在最后的时刻讲述那个青春飞扬的故事的时候，竟然讲得那么辛苦艰难。 这就是那个改变了一个家庭命运的故事：

This is the JC talking about deep learning in Lu Lab.