Sponsorised links
June 2008
An Empirical Evaluation of Deep Architectures on Problems with Many Factors of Variation [PDF]
by ogrisel
(via)
Recently, several learning algorithms relying on models with deep architectures have
been proposed. Though they have demonstrated impressive performance, to date, they
have only been evaluated on relatively simple problems such as digit recognition in a controlled environment, for which many machine
learning algorithms already report reasonable results. Here, we present a series of experiments which indicate that these models show
promise in solving harder learning problems that exhibit many factors of variation. These
models are compared with well-established algorithms such as Support Vector Machines
and single hidden-layer feed-forward neural networks.
YouTube - Visual Perception with Deep Learning
by ogrisel
(via)
A long-term goal of Machine Learning research is to solve highly
complex "intelligent" tasks, such as visual perception auditory
perception, and language understanding. To reach that goal, the ML
community must solve two problems: the Deep Learning Problem, and the
Partition Function Problem.
There is considerable theoretical and empirical evidence that complex
tasks, such as invariant object recognition in vision, require "deep"
architectures, composed of multiple layers of trainable non-linear
modules. The Deep Learning Problem is related to the difficulty of
training such deep architectures.
Several methods have recently been proposed to train (or pre-train)
deep architectures in an unsupervised fashion. Each layer of the deep
architecture is composed of an encoder which computes a feature vector
from the input, and a decoder which reconstructs the input from the
features. A large number of such layers can be stacked and trained
sequentially, thereby learning a deep hierarchy of features with
increasing levels of abstraction. The training of each layer can be
seen as shaping an energy landscape with low valleys around the
training samples and high plateaus everywhere else. Forming these
high plateaus constitute the so-called Partition Function problem.
A particular class of methods for deep energy-based unsupervised
learning will be described that solves the Partition Function problem
by imposing sparsity constraints on the features. The method can learn
multiple levels of sparse and overcomplete representations of
data. When applied to natural image patches, the method produces
hierarchies of filters similar to those found in the mammalian visual
cortex.
An application to category-level object recognition with invariance to
pose and illumination will be described (with a live demo). Another
application to vision-based navigation for off-road mobile robots will
be described (with videos). The system autonomously learns to
discriminate obstacles from traversable areas at long range.
YouTube - The Next Generation of Neural Networks
by ogrisel
(via)
In the 1980's, new learning algorithms for neural networks promised to
solve difficult classification tasks, like speech or object recognition,
by learning many layers of non-linear features. The results were
disappointing for two reasons: There was never enough labeled data to
learn millions of complicated features and the learning was much too slow
in deep neural networks with many layers of features. These problems can
now be overcome by learning one layer of features at a time and by
changing the goal of learning. Instead of trying to predict the labels,
the learning algorithm tries to create a generative model that produces
data which looks just like the unlabeled training data. These new neural
networks outperform other machine learning methods when labeled data is
scarce but unlabeled data is plentiful. An application to very fast
document retrieval will be described.
Sponsorised links
April 2008
ZSFA -- Vellum
by greut
Vellum is a simple build tool like make but written in Python using a simple yet flexible YAML based format. Rather than attempt a full AI engine just to get some software built, I went with the simpler algorithm of a “graph”.
March 2008
Don Quixote Time Series Software
by ogrisel
(via)
Don Quixote is a new business software that uses artificial intelligence and powerful statistical methodology to achieve high forecasting accuracy. No matter if you forecast market shares, sales, profits, demand for services or material, Don Quixote will make your work faster, easier and more accurate and will improve your understanding of the nature of time series.
December 2007
JEliza - Die Opensource KI
by rike_
(via)
Das Computerprogramm JEliza ist die leistungsstärkste Deutsch sprechende künstliche Intelligenz, die den Prinzipien freier Software folgt. Es handelt sich dabei um einen Gesprächssimulator, also eine künstliche Intelligenz, mit der Unterhaltungen ermöglicht werden.
JEliza benutzt ein semantisches Netz, um alle Gesprächsverläufe zu speichern und lernt so dazu.
Robot Powered by Moth’s Brain
by flakki
A new paper studies the effects of robots exhibiting roach-like behaviour on real cockroaches.
Logothèque, en particulier en format vectoriel
by 4004
& 21 others
Logothèque, en particulier en format vectoriel illustrator
November 2007
September 2007
Machine to Transcendent Mind
by YukuanBlog
這本書最合我胃口的是第二章〈小心！前有機器車〉，探討作者對機器自走車的實務經驗。裡面提到作者Hans Moravec在 Mobile Robot Laboratory 接受 Denning Mobile Robotics 委託，研究如何以二十四個聲納組成的障礙偵測裝置，量測、取得的距離資料，完成自主機器車導航的任務。
グーグルのリサーチ担当ディレクター、AIを語る:ニュース - CNET Japan
by kuroyagi
「しかし現在では、Googleとウェブは共同で進化している存在であると理解している。私たちが変更を施すとウェブも変更される。サーチエンジンオプティマイザーは私たちをじっと観察している。そして、私たちが動くとこれらも動く。Googleとウェブとの間の相互作用によってウェブは異なる方向に動くのだ」（Norvig氏）
August 2007
ICML 2007 - PRELIMINARY VIDEOS FROM THE SPOT
by ogrisel
(via)
The 24th Annual International Conference on Machine Learning is being held in conjunction with the 2007 International Conference on Inductive Logic Programming at Oregon State University in Corvallis, Oregon. As a broad subfield of artificial intelligence, machine learning is concerned with the design and development of algorithms and techniques that allow computers to "learn". At a general level, there are two types of learning: inductive, and deductive.
July 2007
Elefant - What is Elefant
by ogrisel
Elefant (Efficient Learning, Large-scale Inference, and Optimisation Toolkit) is an open source library for machine learning
Elefant include modules for many common optimisation problems arising in machine learning and inference. It is designed to be modular and easy to use. Framework provides easy to use python interface, which can be use for quick prototyping and testing inference algorithms.
Amazon.com: "Machine Learning and Its Mathematics, Philosophy & Physics"
by ogrisel
(via)
Interesting reading list on machine learning.
June 2007
Artificial Intelligence: A Modern Approach
by ogrisel
The leading textbook in Artificial Intelligence.
Used in over 1000 universities in 91 countries (over 90% market share).
The 85th most cited publication on Citeseer.
