s*x
发帖数: 8041 | 1 还是icar吧，一定要有augmented reality |
|
d*****u
发帖数: 17243 | 2 任何model都不可能account for all attributes。否则就不是model而是description。
至于说bias，一方面可以考虑prior，另一方面可以增加连接的复杂度（如果需要的话
）。换言之，人不是计算机，所以受记忆等因素的影响很大。但是这些都是基本模型的
augment而已，都是可以数学实现的。而因为这些附加原因就放弃整个model的方法是可
笑的。 |
|
m**t
发帖数: 1292 | 3 Per LightReading article:
Huawei Opens New Jersey Facility
JULY 29, 2011 | Post a comment
BRIDGEWATER, N.J. -- Huawei, a leading global information and communications
technology (ICT) solutions provider, today announced that it will hold a
Ribbon Cutting ceremony on Friday, July 29th to mark the official opening of
its Northeast regional headquarters office in Bridgewater, New Jersey. In
addition to sales, services, and operations to deliver on the company’s
customer-centric promise; the Bridgew... 阅读全帖 |
|
p****t
发帖数: 11416 | 4 著名IT网站PC World今天刊文称，尽管苹果平板电脑iPad的推出令人兴奋，但这款产品
却非常遗憾地缺失了不少功能。如果苹果希望这款产品能够改变世界，那么以下功能是
iPad必不可少的。
1、多任务同时运行
iPad所采用的操作系统不支持多任务同时运行，甚至不支持Safari浏览器同时打开多个
页面。用户无法在收听音乐的同时浏览网页，无法在Facebook和Twitter之间来回转换
，也无法在拨打网络电话的同时编辑文档。
2、Adobe Flash
可以说，不支持Adobe Flash的网络就不能算作真正的网络。iPhone手机不支持Adobe
Flash，用户还可以接受。但对于iPad这样一款屏幕达9.7英寸的产品来讲，不支持
Adobe Flash就太令人遗憾了。毕竟，很多用户希望能够通过iPad观看Hulu上的视频。
3、前置摄像头/视频通话功能
或许用户不会举着平板电脑来拍摄照片，但有些增强实境（augmented reality）应用
会需要这样一个前置摄像头。如果苹果能够在iPad当中加入摄像头和视频通话功能，那
么这款产品将成为拍摄和编辑YouTube视频的理想工具 |
|
|
|
|
b******y
发帖数: 9224 | 8 Hey goodbug, you just nailed it on the head.
I guess the only way to make money out of open source projects is to provide
augmented service.
For example, IDE could be free, but a better design tool or something built
on top of the IDE could charge money, just 2 cents... |
|
n**a
发帖数: 12 | 9 Hello,
Amazon.com is looking for experienced engineers with Machine learning/Data
mining/Hadoop background. Please send your resumes to n******[email protected]
Many positions open, location- Seattle, WA
Job description: SDE
The product catalog is a key business asset and differentiator for Amazon.
The Catalog Quality organization is chartered with the goal of improving the
quality of this data by building systems and employing automated techniques
to identify and fix discrepancies and enrich the dat... 阅读全帖 |
|
p*****s
发帖数: 344 | 10 那是两个应用。
上面的是favorite place，用二维码标识网址，手机就是一读码器。
下面的goggle才是图像识别应用。
再进一步就是终结者那样的visual augmentation。 |
|
c********e
发帖数: 383 | 11 because of static keyword.
1. u can only have pointer or reference or any conbination of the two for ty
pe X as non static member.
2. u can not use non static member as ur constructor augments
what book is this? i would recommend u to read another book |
|
c********e
发帖数: 383 | 12 I didnt read any book nor did i looked up the C++ standard for these rules.
Its all logical conclusion.
1. If u have struct X { X x;} then there is a loop of composition in there.
compiler will not allow this to happen.
2. before the constructor is called the object is not fully constructed, in
ur case, as the augment input, the non-static member is definitely no
t initialized...and u want to use it? |
|
g*********n
发帖数: 43 | 13 Is there any open source code for Minimum spanning tree, connected
components?
Need to use and augment these algorithms to do my own work. Thanks a lot! |
|
c********e
发帖数: 383 | 14 the code is just an example to recreate the problem. never done a template
function that has no augment.
I searched in the gcc release note but didnt find anything that fixed this
after 3.2.3 but apparently something is done here.
thx. |
|
b********h
发帖数: 119 | 15 Does that mean it is both a map and a queue? If so, I guess you can augment
a binary search tree. keep a pointer between the elements pushed into this
data structure to maintain the queue structure. |
|
t********n
发帖数: 611 | 16 Can't run this Gauss elimination program. Got error message : not enough
input arguments.
function x = gauss(A,B)
%The sizes of matrices A,B are supposed to be NA x NA and NA x NB.
%This function solves Ax = B by Gauss elimination algorithm.
NA = size(A,2); [NB1,NB] = size(B);
if NB1 ~= NA, error(’A and B must have compatible dimensions’); end
N = NA + NB; AB = [A(1:NA,1:NA) B(1:NA,1:NB)]; % Augmented matrix
epss = eps*ones(NA,1);
for k = 1:NA
%Scaled Partial Pivoting at AB(k,k) by Eq.(2.2.... 阅读全帖 |
|
d********r
发帖数: 1238 | 17 上个季度居然50% 新增股票都给了RSU，自己涸泽而渔了
http://online.wsj.com/news/articles/SB1000142405270230341710457
Paying workers with shares may not involve cash. But as Twitter TWTR +4.24%
investors are learning, it comes at a real cost.
Shares of the micromessaging platform tumbled Tuesday as a lockup expired on
stock amounting to roughly two-thirds of its diluted share count. Some of
those shares were held by early Twitter investors. Many were held by
employees granted them to augment cash salaries.
The massive e... 阅读全帖 |
|
s*i
发帖数: 5025 | 18 augmented reality，似乎不是现场就是看到3D，而是真是影像混合虚拟影像。 |
|
f********x
发帖数: 99 | 19 Batch is a special case of streaming
Posted on September 15, 2015 by Kostas Tzoumas
In recent blog posts, we introduced what we deem as requirements for systems
to classify as stream processors, and followed up with a detailed
comparison of current approaches to data streaming, including extensive
experiments comparing Apache Flink and Apache Storm.
We are not the only ones to make the point that streaming systems are
reaching a point of maturity which makes older batch systems and
architectures... 阅读全帖 |
|
h**********c
发帖数: 4120 | 20 Last time got a friend here an intern interview, however, the position is
still not finalized due to various reasons.
This position is a little bit higher in requirement.
Thanks for your reading and time. Please PM if you are a strong candidate!
Posting Job Title:
Business Analyst
State:
California
City:
Torrance (Los Angeles, Orange county)
Job Description:
The digital future has arrived and the tax and accounting professions are
changing rapidly. Professionals today have different needs, expec... 阅读全帖 |
|
l*******m
发帖数: 1096 | 21 因为nn不是convex, 收敛性是没法证明。大家都是把logistic regression的sgd的各种
变种拿到nn来. 我估计在lr情况下，这个片文章的方法应该是ok.
ensemble可以用几个不同模型，也可以用一个模型，不同的data augmentations. 比如
鉴别植物，可以用原始图片infer, 再把图片转个角度再infer一次，做个平均 |
|
|
S***s
发帖数: 104 | 23 就是排除AI augmented Human的可能性
新纪元了,不管是哪一家的 |
|
|
E*********g
发帖数: 185 | 25 码农终于开发出了上吊的绳子
https://arxiv.org/abs/1709.05703
AI Programmer: Autonomously Creating Software Programs Using Genetic
Algorithms
Kory Becker, Justin Gottschlich
(Submitted on 17 Sep 2017)
In this paper, we present the first-of-its-kind machine learning (ML) system
, called AI Programmer, that can automatically generate full software
programs requiring only minimal human guidance. At its core, AI Programmer
uses genetic algorithms (GA) coupled with a tightly constrained programming
language that mi... 阅读全帖 |
|
w***g
发帖数: 5958 | 26 如果是几十上百个词的短文本，我估计rnn不会有提高（NLP专家请指正）。你有时间还
不如试试data augmentation, 比如随机增减几个词，同义词替换啥的。不过你现在最
重要的是establish问题复杂度，就是上次说的对比不同人的区别，这样知道什么时候
停止优化。不然老板拍脑袋定指标，终有一天会把你逼死。
：正在测试CNN，epoch还没训练完，不过目前的accuracy已经达到70%，
： |
|
w***g
发帖数: 5958 | 27 如果是几十上百个词的短文本，我估计rnn不会有提高（NLP专家请指正）。你有时间还
不如试试data augmentation, 比如随机增减几个词，同义词替换啥的。不过你现在最
重要的是establish问题复杂度，就是上次说的对比不同人的区别，这样知道什么时候
停止优化。不然老板拍脑袋定指标，终有一天会把你逼死。
：正在测试CNN，epoch还没训练完，不过目前的accuracy已经达到70%，
： |
|
n******7
发帖数: 12463 | 28 看了一下，64张图训练太少了吧
估计即使transfer learning目前也不行
另外，我感觉可以把data augmentation看作学习的一步，以后可能可以更好整到学习
过程中
毕竟DNN是个复杂的结构，没有足够的data怎么train？ |
|
w***g
发帖数: 5958 | 29 Raid1更快，多个盘镜像提速。走pcie的m2盘比ssd又快不少，就是堆不了太多还占pcie
。估计cpu做augmentation会先到瓶颈。
：如果用ssd再组个raid0，没意义么？
： |
|
w***g
发帖数: 5958 | 30 最近看到了imgaug，信心崩塌了。
之前做的时候还有torch，还指望R，matlab啥的也可能都要上DL。
现在基本上尘埃落定DL训练非python不可了。
下一步就是把augmentation从picpac中刨掉了。
剩下的C++代码都是IO intensive，其实用python写也不会太慢。
得想点有意思的东西做做才行。 |
|
L****8
发帖数: 3938 | 31 I got it from recent experiments using unet |
|
|
g*******u
发帖数: 3948 | 33 这个本质不就是让样本更丰富吗？ 让机器看到更多可能的 然后就会学的好点？
这个一直用的吧 尤其数据不太多的时候 |
|
L****8
发帖数: 3938 | 34 very useful if the objects have nonlinear deformation
such as heart in medical images |
|
