|
w*********y
发帖数: 7895 | 2 这里并没有引用任何文献，只是基于我对自己孩子的观察。
在我家娃第一次玩PUZZLE的时候，我就发现2个人的learning approach
很不一样。小肥靠观察和记忆，所以会倒着拼图。小肉靠learning by trial. 一开始
的时候，小肥表现出很快的学习能力，一下就学会了，而且可以纠正你拼错的地方。经
常指挥小肉怎么拼。小肉则还在自己折腾，常常会放错。到现在，她们已经玩了很多
PUZZLES了，小肉这个学的慢的娃，赶上了小肥，最早学会拼48片图，小肥则常常要父
母帮忙一起拼才会拼。这个时候我意识到我对着两种learning approaches有了理解上
的偏差。正因为小肉的learning by trial的学习方法，所以她有了更多的练习机会，
所以知道自己的errors在什么地方。而且练习了，就learning by trial的速度加快了
，因为可以判断某个地方是否需要尝试。小肥因为依赖父母，要看父母拼，记忆图片，
碰到新的图片时候无法自己下手。所以2个方法应该有各自优点缺点吧。 |
|
a****y
发帖数: 255 | 3 我感觉,美国学校训练好的小孩,是非常可能learning by trial,因为这更接近西方人的
思维方式;learning by memory更接近东方人(东亚人)的思维方式;
这两种方式各有优点,正如LZ所说,开始时,learning by memory学习的快;但时间长了,
learning by trial的威力可以显现出来.其实,这体现的是learning approach的不同,
也体现了孩子对大人的依赖程度的不同,就是哪个小孩更独立.LZ可以看看,小肉是不是
平时更独立一些? |
|
发帖数: 1 | 4 Postdoctoral Fellowship-Machine Learning & Health Analytics in the
Department of Computer Science and Engineering at the University of
Connecticut
About University of Connecticut:
The University of Connecticut (UConn) was founded in 1881, and is ranked
within the Top 20 Public Universities and within Top 60 National
Universities in the United States and is amongst the elite institutions
designated as “highest research activity” by the Carnegie Classification
of Institutions of Higher Education. ... 阅读全帖 |
|
r*****e
发帖数: 7853 | 5 这是我最不喜欢有人搞dnn的地方，什么feature也不提取，learn出来的是一坨屎
比如语音识别，不用feature learn的狗屁不是，用恰当的feature就用不上几个layer
的dnn了，一般的machine learning就行了
我不否认machine learning的用处，但现在满大街的dnn，扯淡的多
[在 CRH1235 (江左没狼) 的大作中提到：]
:Deep learning的好处是不需要人工提取feature。
:最简单的用收盘价就可以学习了，那input就是个一维（ｎｘ１）的vector（ｎ是时间
:），当然这样效果不一定好。讲究点的可以把开盘，收盘，高点，低点，成交量等等
一起当输入，这样输入就是（ｎｘ５）的二位matrix。还有其他的也可以加进去，输入
就是（ｎｘｍ），但是还是二维的。 |
|
I***a
发帖数: 13467 | 6 【 以下文字转载自 Immigration 讨论区 】
发信人: bigpapa (bigpapa), 信区: Immigration
标 题: 求审稿机会： machine learning, deep learning, vision, etc
关键字: 求审稿机会，
发信站: BBS 未名空间站 (Mon Nov 21 17:51:40 2016, 美东)
如题，求machine learning, 相关方面审稿，phd 做的是ML 和 vision，NLP。
目前 research scientist 在做deep learning。
如果有机会，请发站内信，可以提供简历，审稿又快又好。 |
|
C**********r
发帖数: 1032 | 7 Education Hill Learning Center: A Love & Joy Bilingual (English/Chinese)
Daycare
Education Hill Learning Center, a newly opened bilingual
(English/Chinese), Licensed daycare, now has multiple opening.
At Education Hill Learning Center we advocate the key to long-term
educational success is making sure children get the best start possible.
We emphasize the importance of the early childhood education through
play. We focus on not only academics and physical skills development but
also social, emot... 阅读全帖 |
|
r********k
发帖数: 170 | 8 -
I'd highly recommend this family care to parents. Teachers take good care of
kids. I was worried that my son would get sick quite often in the first few
months. It turned out that he only caught a cold and recovered quickly. It
only took him two weeks to adjust to the new environment and now he goes to
the center happily. Most importantly, the owner Janet loves her job. She
likes children and is experienced in teaching little kids. My son learned to
eat with spoon several days after he went to... 阅读全帖 |
|
l***o
发帖数: 1111 | 9 By HOWARD BECK
The next inevitable phase of the Jeremy Lin story is under way, in which
reasonable doubts and manufactured controversy replace the initial burst of
elation. This is where the mythology fades and the fundamental work for the
Knicks begins.
This is the point when Lin must make the most critical adjustment: from
instant celebrity to enduring basketball star.
Since the Lin-fueled seven-game winning streak that saved the season, the
Knicks have lost four of seven games, while Lin has ... 阅读全帖 |
|
l**7
发帖数: 3940 | 10 wow, Li Na 貌似被捧为前三值得学习的明星了，莎娃和A拉还没上市呢
1. Learn to Play Like: Serena Williams
2. Learn to Play Like: Victoria Azarenka
3. Learn to Play Like: Li Na
4. Learn to Play Like: Maria Sharapova (coming May 7)
5. Learn to Play Like: Agnieszka Radwanska (coming May 14) |
|
g*****h
发帖数: 15 | 11 Math Learning Ebooks Collection
English | 2011 | PDF | 700MB
130 Math Learning Ebooks Collection,Books List:
Algebra
Abstract Algebra
Algebra Groups
Algebra Challenging Problems
Algebraic groups and Discontinuous Subgroups
Elementary Linear Algebra
Elements of Abstract and Linear Algebra
Feynman Physics Lectures
Lectures on Matricies
Linear Algebra
Math Resource Part I Algebra
Complex Algebra 2001
Fundamental Problems in Algorithmic Algebra
Workbook in Higher Algebra
Basic Math and Algerba
Matri... 阅读全帖 |
|
|
|
y*******y
发帖数: 46 | 14 背景： 本科经济学学士，国际商务英语专业，来美13年，Top50大学，工业关系和人力
资源 MS。工作11年，现就职小型非营利培训与研究机构，研究部主任，研究主要方向
在交通行业内frontline technical worker training, apprenticeship, career
pathways and career ladders, credentialing/certification/qualification,
learning outcomes assessments, cost-benefits analysis, training ROI, etc. 因
为是针对一个行业，而且研究侧重实践性，文章主要发在本行业内的Journals和
conference proceedings。
做这个职位也三四年了，想回学校充实一下理论基础，结合过去十几年工作中遇到的问
题，做一些有一定影响力的研究。准备读在职，希望是Top 10学校（因为有on-site课
程，只能在家附近找），公司可以资助一部分或全部（一年一两万吧），而且通过公司
可以接触到一些研究需要的数据等... 阅读全帖 |
|
l******n
发帖数: 1250 | 15 看了这么多quant的工作，觉得machine learning是必备的。
但是关于machine learning的深度和广度，不知道哪一个更重要
CS PHD的machine learning，往往比较注重理论。他们可能几年就搞SVM，或者Neural
Network，搞出很多很多变体来。
向我们EE的PHD，也做machine learning，但是往往比较注重实践。我们是尝试不同的
学习方法，比如SVM， Neural Network，adaboost,random forest, 用在不同的领域，
比如分类，预测等等，然后比较优略。
我不知道工业界对machine learning，是注重深度多一些，还是广度？ |
|
t*o
发帖数: 2606 | 16 Fisher-Price Laugh & Learn Learning Farm 只要 $40 (50% off 啊)
link:
http://cribeasy.com/go/walmart/11090818
**** 这个deal 非常好，快买吧。****
相关推荐说明:
icesoar:
这个玩具小孩很喜欢。不会走的时候，可以绕着玩具转圈，可以钻来钻去。学走路了，
就扶着玩具走来走去，还会捉迷藏。会走路了，就开始自己顽皮的玩玩具上的mini
activities了。带来了无限的欢乐呀。
要是你家里够大，可以收留着，每个玩具都有着很美好的回忆。
christol8888:
我表妹买了learning home跟learning kitchan,两个小孩玩遍3个，比较之下，还是觉
得learning farm最多东西玩。
=============================
今天其他推荐：
[0] amazon $5 off $25 好 deal， 不可错过阿
注意把那下面 “sign me up to receive Amazon's ..." 前面的对勾去掉，否则他们
会经... 阅读全帖 |
|
w*******y
发帖数: 60932 | 17 Tag Learn to Read Super Book Pack
Ages 4-7 Years For Only $34.99 from retail of $59.97!
I search and didn't find this deal.
Link: Link:
http://shop.leapfrog.com/leapfrog/browse/productDetail.jsp?productId=TAG63003
Includes Tag Learn to Read Book Set 1 (Short Vowels), Learn to Read Book Set
2 (Long Vowels, Silent E and Y) and Learn to Read Book Set 3 (Consonants).
For readers who have just begun, this book pack brings three times the fun!
Perfectly sized for little hands, the Tag Reader is a touc... 阅读全帖 |
|
w*******y
发帖数: 60932 | 18 For free shipping use 2011SELECTGUEST or alternate 2011SIGNATUREGUEST
***shipping discount shows on credit card info page***
Fisher-PriceLaugh 'n Learn Learning Workbench:
http://www.bonton.com/shop/kids/toys/classic-toys/fisher-price-
$13 shipped
Three interactive learning modes and lots of tool friends bring the
workbench to life; with 15+ sing-along songs and tunes, fun sounds and
phrases. Theres a dancing drill, spinning vise, twirling paint roller, count
-and-slide ruler, and rattling ham... 阅读全帖 |
|
w*******y
发帖数: 60932 | 19 LeapFrog Learn and Groove Music Wand now $8 on Amazon:
http://www.amazon.com/LeapFrog-Learn-Groove-Music-Wand/dp/B0038
As children move the wand, they can change the music's tempo and add new
sounds, creating their own masterpieces. Parents can connect to the online
LeapFrog Learning Path for customized learning insights and ideas to expand
the learning. Appropriate for children ages 3 years and up.
|
|
|
G*******s
发帖数: 10605 | 21 【 以下文字转载自 SmartShopper 俱乐部 】
发信人: GoRockets (火箭加油), 信区: SmartShopper
标 题: Fisher-Price Laugh & Learn Learning Farm $39.97
发信站: BBS 未名空间站 (Wed Dec 5 19:52:09 2012, 美东)
这个我们家有，非常好，儿歌很多，我们家三岁多了还经常玩一玩
Walmart has Fisher-Price Laugh & Learn Learning Farm for $39.97 (this is
almost 50% off original price) with free store pickup. This is a very good
set for babies up to 3 years old. Amazon still has it for $70+.
http://userquote.com/forum/index.php?topic=1649.msg1930#msg1930 |
|
G*******s
发帖数: 10605 | 22 【 以下文字转载自 SmartShopper 俱乐部 】
发信人: GoRockets (火箭加油), 信区: SmartShopper
标 题: Fisher-Price Laugh & Learn Learning Farm $40
发信站: BBS 未名空间站 (Sat Jan 19 08:40:49 2013, 美东)
Walmart has Fisher-Price Laugh & Learn Learning Farm for $40 (this is almost
50% off original price) with free store pickup. This is a very good set for
babies up to 3 years old. Amazon still has it for $60+.
http://userquote.com/forum/index.php?topic=1716.msg1998#msg1998 |
|
v******s
发帖数: 272 | 23 What Obama And America Should Learn From China
Shaun Rein, 01.24.11, 12:15 PM EST
There are ways in which Hu Jintao is being wiser than us.
There has been much discussion lately of how China survived the
financial crisis much more robustly than the United States. Many credit
Beijing's centralized power structure, which Sen. Harry Reid mistakenly
(and foolishly) calls a dictatorship under President Hu Jintao.
Political scientist Francis Fukuyama recently wrote that China's command
system allows f... 阅读全帖 |
|
j*****y
发帖数: 549 | 24 【 以下文字转载自 SanFrancisco 讨论区 】
发信人: jojojoy (Jojo), 信区: SanFrancisco
标 题: 替朋友转个learning home （san jose地区），30刀
发信站: BBS 未名空间站 (Tue Apr 15 20:22:19 2014, 美东)
帮朋友转一个learning home，就是经典的Fisher-Price Laugh & Learn Learning
Home Palyset，八成新吧，零件都在。这个玩具的好我就不说了，6个月娃就可以玩了
，可以玩到几岁。30刀。san jose地区可以自取，感兴趣的mm可以站短我，谢谢 |
|
h********y
发帖数: 43 | 25 大家好，想来想去就只有这个版块偏学术点，如有不妥，请指出。
RT，我最近接到两个postdoc的offer：一个是machine learning，研究里算法/理论和
应用各占一半一半吧。另一个是bioinformatics，研究是用计算模型和machine
learning算法来分析大量且高维的genomic data。
这两个offer的location和学校排名都不相上下，我自己打算走academia最后申教职。
我本身是machine learning背景出身，对bioinformatics从来没接触过。
想在这里请问一下大家，征询下意见：
1. 这两个方向具体未来发展前景怎样？我没做过bioinformatics，对这个领域的发展
不太清楚，请大家不吝赐教。
2. 我自己本身对证machine learning的理论类的有点做烦的感觉（应用类没烦）。但
在这几日看bioinformatics的时候，觉得更偏应用一些，觉得如果纯做应用会不会对未
来发展不大好，大家觉得呢？
谢谢大家！
Update: 我现在主要纠结选哪个方向，请大家指点一二，多谢！ |
|
s******g
发帖数: 67 | 26 My daughter has been attending Education hill learning center since 2011.
She enjoys everyday and makes good friends there. I think three things make
Education hill learning center a special place for young children. First,
teachers teach different and appropriate materials to kids based on their
ages and abilities. For example, my daughter is good at learning new words
and counting numbers. The teachers realized that and taught her a lot of
words. By the age of 3, my daughter can read about 80 ... 阅读全帖 |
|
j*****y
发帖数: 549 | 27 【 以下文字转载自 SanFrancisco 讨论区 】
发信人: jojojoy (Jojo), 信区: SanFrancisco
标 题: 替朋友转个learning home （san jose地区），30刀
发信站: BBS 未名空间站 (Tue Apr 15 20:22:19 2014, 美东)
帮朋友转一个learning home，就是经典的Fisher-Price Laugh & Learn Learning
Home Palyset，八成新吧，零件都在。这个玩具的好我就不说了，6个月娃就可以玩了
，可以玩到几岁。30刀。san jose地区可以自取，感兴趣的mm可以站短我，谢谢 |
|
x***n
发帖数: 70 | 28 我是Biomedical Engineering的PHD，年底毕业。做了一些生理信号分析的项目，
machine learning的技术用的较多，有几篇相关的论文（但只是对machine learning的
应用）。我想问下machine learning的就业机会怎么样？据我所知，它的工作要求似乎
很偏重统计。我的编程能力还不错，本科是EE出身，java和C++都用的很多。想请板上
前辈帮忙分析下，我如果想以machine learning这一块作为目标，应该注意哪些？
谢谢！ |
|
f*********m
发帖数: 726 | 29 有不少科技公司中有用machine learning/data mining 做business analysis，可能属
于Marketing或者Operation部门。
不知到这种部门做machine learning/data mining有前途吗，会不会用的machine
learning/data mining不太深，学不到太多东西，出来也不好再做要求比较高的
machine learning/data mining工作了？ |
|
f*********m
发帖数: 726 | 30 有不少科技公司中有用machine learning/data mining 做business analysis，可能属
于Marketing或者Operation部门。
不知到这种部门做machine learning/data mining有前途吗，会不会用的machine
learning/data mining不太深，学不到太多东西，出来也不好再做要求比较高的
machine learning/data mining工作了？ |
|
z**l
发帖数: 113 | 31 【 以下文字转载自 JobMarket 讨论区 】
发信人: zzxl (I love Brownie), 信区: JobMarket
标 题: Cheminformatics intern (Machine learning focus)
发信站: BBS 未名空间站 (Tue Mar 11 11:22:57 2014, 美东)
我们组要招一位实习生，主要是与chemists/biologists合作build一些predictive
models.贴的有些晚了，不过希望感兴趣的在本周把简历递到DTomandl [at] Dow.com。
可以抄送一份简历到baolanbujing[at]gmail.com. 有问题可以直接问我。
Cheminformatics Summer Intern 2014 at Dow AgroSciences
Dow AgroSciences seeks one summer intern for the Cheminformatics group
within the R&D Data Analysis group. Send CV to DToma... 阅读全帖 |
|
m********u
发帖数: 3942 | 32 在硅谷，有兴趣请站内或者邮件 [email protected]
Education:
MS/Ph.D in computer science or software engineering or equivalent
Key Qualifications:
Relevant professional experience (3+ years) in machine learning, data mining
or equivalent fields
Familiar with large data and distributive system
Comfortable to code in one or more in following: Java / C++ / Python
Self-driven and diligent for solving real world problems
Preferred Requirements:
Experience/knowledge of large scale machine learning applied to very... 阅读全帖 |
|
z*******h
发帖数: 346 | 33 AI包括乱七八糟的东西，machine learning是一部分。不过以前AI名声坏掉的时候，
machine learning就叫machine learning, 现在叫AI显得牛逼。Datamining和machine
learning有很大交集，基本上是喜欢叫啥就叫啥。 |
|
l*****n
发帖数: 1763 | 34 我家宝宝原来最喜欢Fisher Price Learning Home，从8个月一直玩到一岁出头。天天
玩。
很犹豫地买了Fisher Price Learning Farm，担心太相似了。结果从一岁半玩到现在（
刚刚两岁），很喜欢。一来歌多（比Fisher Price Learning Home），二来有动物（大
一点才喜欢动物）。而Fisher Price Learning Home则很少玩了。btw，我家玩具极多
。这个还是很喜欢的。
当然每个孩子的兴趣都不尽相同。仅供参考。 |
|
j*****y
发帖数: 549 | 35 【 以下文字转载自 SanFrancisco 讨论区 】
发信人: jojojoy (Jojo), 信区: SanFrancisco
标 题: 替朋友转个learning home （san jose地区），30刀
发信站: BBS 未名空间站 (Tue Apr 15 20:22:19 2014, 美东)
帮朋友转一个learning home，就是经典的Fisher-Price Laugh & Learn Learning
Home Palyset，八成新吧，零件都在。这个玩具的好我就不说了，6个月娃就可以玩了
，可以玩到几岁。30刀。san jose地区可以自取，感兴趣的mm可以站短我，谢谢 |
|
Y*****2
发帖数: 38613 | 36 we do not spontaneously learn that we dont' learn what we don't learn
is better. |
|
C**********r
发帖数: 1032 | 37 Education Hill Learning Center enrolling now
Education Hill Learning Center, a newly opened bilingual daycare, now
has multiple opening.
Education Hill Learning Center provides full-time and part-time caring
for children ages 4 months to 12 years. Based on developmental theories
and child-centered methodologies, we offer children a play-based program
which focuses on the emotional and social development and which promotes
their physical, cognitive, creative growth. Programs include activities
t... 阅读全帖 |
|
n**a
发帖数: 12 | 38 Hello,
Amazon.com is looking for experienced Research Scientist with Machine
Learning, Modelling, Analytics background. Please send your resumes to
n******[email protected]
Many positions open, location- Seattle, WA
Job description: Research Scientist
Product Ads is breaking fresh ground, pioneering a crucial platform for
future Amazon growth. This is high-profile business unit, with support and
interest from all parts of Amazon and top management. We are entrepreneurial
and have a basis for action w... 阅读全帖 |
|
l*****9
发帖数: 9501 | 39 Jeremy Lin Must Learn from Derrick Rose and Lead by Example By Ben Shapiro(
Featured Columnist) on March 12, 2012
If you read the above headline and think this is going to be a column in
which Jeremy Lin is compared to Derrick Rose, think again.
Derrick Rose is either the league's best player or one of the top five, take
your pick and debate amongst yourselves.
But there is one arena where the two players intersect. That would be
leadership.
The Bulls don't win just because of Rose's outstanding... 阅读全帖 |
|
a*****g
发帖数: 19398 | 40 McCormick Elementary Students Learn Go (And Math)
Students at McCormick Elementary, in Chicago, IL, had the opportunity recent
ly to learn to play go from Xinming Simon Guo 2d, a licensed math teacher an
d founder of the GoAndMath Academy. “Students were playing a simple game du
ring the class, blissfully unaware that they were also working on math skill
s as they put every stone on the board and counted the result at the end of
the game,” Guo told the E-Journal.
At McCormick, the go class is pa... 阅读全帖 |
|
w*******d
发帖数: 59 | 41 有句话叫“半桶水响叮当”……今天终于见识到了……Sutton和barto的书老掉
牙这种话话也真敢说得出口=_=……那大概图灵的paper是不是也可以扔垃圾箱里了。
MCTS作为算法本身和RL是两个独立概念。当RL里的value和policy function固定不变的
情况下，这个数学模型就退化为统计上的Markov decision process。而MDP最优决策是
一个NP-hard问题，所以利用MCTS算法可以近似的给每步搜索最优解。
当value和policy不知道的情况下，你需要通过MDP里不断的数据反馈去学习这两个函数
，这个过程叫做reinforcement learning。换句话说，我可以不用MCTS去在每步寻找最
优决策，而换另一种搜索方法，但是这个过程依然是reinforcement learning……不知
道这样讲清楚了没有……
Google的最大贡献就是用deep belief nets来model这两个函数+MCTS搜索。这个被
Google称为deep reinforcement learning。换言之，我可以用random forest来mode... 阅读全帖 |
|
G****e
发帖数: 11198 | 42 Learn from yesterday, live for today, Hoke for tomorrow - JNW
http://mgoblog.com/diaries/learn-yesterday-live-today-hoke-tomo
By Lordfoul — October 9th, 2011 at 11:54 AM — 17 comments
Filed under: 6-0 is rare Al Borges is a mad magician football Greg Mattison
is Gary Moeller++ Junior Denard overcomes his mistakes Senior Junior helps
out
“Learn from yesterday, live for today, hope for tomorrow. The important
thing is not to stop questioning.”
Albert Einstein
Learn from yesterday...
6-0 starts for... 阅读全帖 |
|
C**********r
发帖数: 1032 | 43 Education Hill Learning Center enrolling now
Education Hill Learning Center, a newly opened bilingual daycare, now
has multiple opening.
Education Hill Learning Center provides full-time and part-time caring
for children ages 4 months to 12 years. Based on developmental theories
and child-centered methodologies, we offer children a play-based program
which focuses on the emotional and social development and which promotes
their physical, cognitive, creative growth. Programs include activities
t... 阅读全帖 |
|
p********g
发帖数: 84 | 44 I've learned-
that you cannot make someone love you.
All you can do is be someone who can be loved.
The rest is up to them.
I've learned-
that no matter how much I care,
some people just don't care back.
I've learned-
that it takes years to build up trust,
and only seconds to destroy it.
I've learned-
that it's not what you have in your life
but who you have in your life that counts.
I've le |
|
t***r
发帖数: 225 | 45 Title: PostDoc Fellow
Company/Institution: National Institutes of Health Clinical Center
Location: Bethesda, MD 20892, USA
Department: Department of Radiology and Imaging Science
Focus: Large Scale Medical Image Informatics Mining via Deep Learning
Description: A post-doctoral fellowship is available for 2D/2.5D/3D/4D
radiology image processing at National Institutes of Health Clinical Center.
Specific interest areas are image segmentation, modeling, visualization,
pattern recognition, comput... 阅读全帖 |
|
t***r
发帖数: 225 | 46 Title: PostDoc Fellow
Company/Institution: National Institutes of Health Clinical Center
Location: Bethesda, MD 20892, USA
Department: Department of Radiology and Imaging Science
Focus: Large Scale Medical Image Informatics Mining via Deep Learning
Description: A post-doctoral fellowship is available for 2D/2.5D/3D/4D
radiology image processing at National Institutes of Health Clinical Center.
Specific interest areas are image segmentation, modeling, visualization,
pattern recognition, comput... 阅读全帖 |
|
s*******n
发帖数: 9 | 47 machine learning就是让计算机具有学习的能力。算是AI里的一个分支吧。
具体的一些方法包括，concept learning,decision tree learning，Bayes learning,
neural networks,…… |
|
I**********s
发帖数: 441 | 48 In supervised learning, you know the answer and use the answer to direct the
adjustment of parameters to be learned while training.
In unsupervised learning, you don't know the answer. The program learns the
parameters by itself upon the internal relationships of the training data. |
|
B*****l
发帖数: 192 | 49 ls推荐的都不靠谱，不适合做“基本的，textbook类型的”。
如果时间有限，不想花太多时间在公式上，推荐
Introduction to Machine Learning (Adaptive Computation and Machine Learning)
by Ethem Alpaydin
如果想稍微深入了解一下machine learning theory，推荐
Pattern Recognition and Machine Learning (Information Science and Statistics
) by Christopher M. Bishop
对入门的兄弟姐妹来说，这两本读起来要容易多了、舒服多了。 |
|
r********3
发帖数: 2998 | 50 vision水的原因不是在于数学理论和推导不够深入，而是在于没有实际应用，不值得搞
那么多conference，录取那么多paper。vision这种脱离实际的学科，应该把ICCV,CVPR
,ECCV合并到一起，一年录取个60,70篇paper足够了。
要做learning的，就去投真正的learning的。要做math就去投真正的math的。vision这
个一年缩减到60,70 top conf papers，如果还有人一年4篇，那谁也不会说他水了。
另外，做vision和learning的人真的应该好好学习别人做system的。别人好几个人，花1,2年时间做一个system出来。甚至还在production server上测试几个月，然后才投一篇paper。而vision和learning的很多paper就是做一道数学作业。几十行matlab，download一个数据，run一下就得出结果了。
的 |
|
