=人工智能简介= 人工智能 ==
人工智能,是指计算机系统具备从听说读写到搜索、推理、决策和回答问题等类人智能的能力。即,感知、理解、决策的能力。人工智能(Artificial Intelligence),是指计算机系统具备从听说读写到搜索、推理、决策和回答问题等类人智能的能力,即感知、理解、决策的能力。
===人工智能历史===
过去经历了2次高潮与2次低谷
基于大数据的机器学习的算法进步
阅读材料:=== 四个层面 ===
# The rebirth of neural networks-ISCA-2010 [http://pages.saclay.inria.fr/olivier.temam/homepage/ISCA2010web.pdf rebirth_NN]* 目标与功能
==四个层面==: 语音识别、机器视觉、自然语言理解: 智能问答是综合以上功能的高级系统
===目标与功能分类===* 核心技术
语音识别、机器视觉、自然语言理解。 : 特定算法、机器学习算法、深度神经网络
智能问答是综合以上功能的高级系统。* 软件工具
===核心技术分类===: TensorFlow / Caffe / Torch
特定算法 机器学习算法 深度神经网络* 底层硬件
===软件工具===TensorFlow : 可编程逻辑阵列 FPGA / Caffe 通用图形处理器 GPGPU / Torch通用处理器 CPU 群集
===底层硬件分类国际研究 ===
可编程逻辑阵列 FPGA [http:/ 通用图形处理器 GPGPU / 通用处理器 CPU 群集research.google.com/teams/brain/ Google Brain]([http://research.google.com/pubs/jeff.html Jeffrey Dean])
==国际研究== [httphttps://research.googlefacebook.com/teams/brain/ Googel_Brainai Facebook AI Research (FAIR)]([http://researchyann.googlelecun.com/pubs/jeff.html Jeff Dean]) [https://research.facebook.com/ai Facebook_AI-Research](Yann LeCun])
[https://www.microsoft.com/en-us/research/group/dltc/ MSR Deep Learning Technology Center (DLTC)]
([https://www.microsoft.com/en-us/research/people/deng/ Li Deng])
[https://www.openai.com/blog/ OpenAI]
([http://www.cs.toronto.edu/~ilya/ Ilya Sutskever]) == 机器学习 == 机器学习(Machine Learning),是指机器从数据中自动分析获得规律,并利用规律对未知数据进行预测。 === 阅读材料 === # Jordan, M. I., and T. M. Mitchell. "Machine learning: Trends, perspectives, and prospects." Science 349, no. 6245 (2015): 255-260. [http://science.sciencemag.org/content/349/6245/255 Machine_Learning_Science_2015]
==机器学习= 工具 ===
Machine Learning'''Python'''
[http://scikit-learn.org scikit-learn]
([https://github.com/scikit-learn/scikit-learn Source Code])
# Jordan, M. I., and T. M. Mitchell. "Machine learning: Trends, perspectives, and prospects." Science 349, no. 6245 (2015): 255-260. [http://science.sciencemag.org/content/349/6245/255 Machine_learning_science_2015]== 深度学习 ==
=深度学习=深度学习(Deep Learning),机器学习中一种基于对数据进行表征学习的方法,试图使用包含复杂结构或由多重非线性变换构成的多个处理层对数据进行高层抽象的算法。
Deep Neural Networks,简称DNN=== 神经网络 ===
深度神经网络,Deep Neural Networks,简称DNN [[卷积神经网络]],Convolutional Neural Networks,简称CNN 历史:The rebirth of neural networks, ISCA 2010.[http://pages.saclay.inria.fr/olivier.temam/homepage/ISCA2010web.pdf Rebirth_NN] === 阅读材料 ===
[[深度学习-入门导读]]
==深度学习工具= 工具 ===
===谷歌==='''Google'''
[https://www.tensorflow.org/ TensorFlow]([https://github.com/tensorflow/tensorflow TensorFlowSource Code])
[http://download.tensorflow.org/paper/whitepaper2015.pdf Google_TensorFlow_whitepaperTensorFlow_Whitepaper]
==='''Facebook==='''
[https://github.com/facebook/fbcunn fbcunn]
[http://torch.ch/ Torch]
([https://github.com/torch/torch7 Source Code])
===微软=== [httphttps://cntkgithub.ai CNTK_Microsfotcom/facebook/fbcunn fbcunn]
[https://www.microsoft.com/en-us/research/product/cognitive-toolkit/tutorials/ CNTK]'''Microsoft'''
===DMLC===[http://cntk.ai CNTK]([https://github.com/microsoft/cntk Source Code])
'''[http://dmlc.ml/ dmlcDMLC]'''
[http://mxnet.io/ MXNet]
([https://github.com/dmlc/mxnet dmlc_mxnetSource Code])
==='''Université de Montréal==='''
[http://www.deeplearning.net/software/theano/ Theano]
([https://github.com/Theano/Theano/ Source Code])
[https://github.com/Theano/Theano/ Theano_code]== 增强学习 ==
增强学习(Reinforcement Learning),是机器学习中的一个领域,强调如何基于环境而行动,以取得最大化的预期利益。 === 阅读材料 ==增强学习=增强学习(Reinforcement Learning)
[[增强学习-入门导读]]
===工具=====谷歌===deepmind/lab '''Google''' [https://github.com/deepmind/lab deepmind_labDeepMind Lab] '''OpenAI'''
===OpenAI===openai[https://universe.openai.com/ OpenAI Universe]([https://github.com/openai/universe openai_universeSource Code])
==机器感知==
机器感知(Machine Perception),如语音,图像,视频,手势,姿态等
==<B>以下重点讨论'''基于深度学习的机器感知</B>=='''
===语音识别===
语音识别(Automatic Speech Recognition),简称ASR
基本工具
*:Long short term memory neural network (LSTM)
:#Long short term memory neural computation, Neural computation 9 (8), 1735-1780, 1997. [http://ieeexplore.ieee.org/document/6795963 LSTM]
*:Connectionist temporal classification Long short term memory neural network (CTCLSTM) :#Connectionist temporal classification: labelling unsegmented sequence data with recurrent Long short term memory neural networkscomputation, ICML 2006Neural computation 9 (8), 1735-1780, 1997. [http://ieeexplore.ieee.org/document/6795963 LSTM]
*:Gated Recursive Unit Connectionist temporal classification (GRUCTC):#On the Properties of Neural Machine TranslationConnectionist temporal classification: Encoder-Decoder Approacheslabelling unsegmented sequence data with recurrent neural networks, SSST-8, 2014ICML 2006.
Alex Graves,DeepMind研究员,语音识别多项技术开创者。详见Google Scholar *: Gated Recursive Unit (GRU):# On the Properties of Neural Machine Translation: Encoder-Decoder Approaches, SSST-8, 2014. [httpshttp://scholarwww.googlecs.comtoronto.hkedu/~graves/citations?user=DaFHynwAAAAJ Alex Graves] ,DeepMind研究员,语音识别多项技术开创者。 :#Towards End-To-End Speech Recognition with Recurrent Neural Networks, ICML 2014.:#Speech recognition with deep recurrent neural networks, 2013.:#Hybrid speech recognition with deep bidirectional LSTM, ASRU 2013.:#Connectionist temporal classification: labelling unsegmented sequence data with recurrent neural networks, ICML 2006.
Google Speech
:# Google Speech Processing from Mobile to Farfield, CHiME 2016. [http://spandh.dcs.shef.ac.uk/chime_workshop/presentations/CHiME_2016_Bacchiani_keynote.pdf Google_Speech_Processing]
===计算机视觉===
计算机视觉(Computer Vision),简称 CV
* Object Detection
Ross Girshick, FAIR研究员,R-CNN的开创者。 [httpshttp://scholarwww.googlerossgirshick.com.hkinfo/citations?user=W8VIEZgAAAAJ Ross Girshick],FAIR研究员,R-CNN算法的开创者。
:<B>R-CNN (Region-based Convolutional Network method)</B>
:<B>Faster R-CNN (Faster Region-based Convolutional Network method)</B>
::#Faster R-CNN Towards real-time object detection with region proposal networks, NIPS, 2015.
::• R-CNN(Matlab): https://github.com/rbgirshick/rcnn
::• Fast_R-CNN(Python): https://github.com/rbgirshick/fast-rcnn
::• Faster_R-CNN(matlabMatlab): https://github.com/ShaoqingRen/faster_rcnn
::• Faster_R-CNN(Python): https://github.com/rbgirshick/py-faster-rcnn
==机器认知==
机器认知(Machine Cognition),自然语言理解,推理,注意,知识,学习,决策,交互等。Cognition),自然语言理解、推理、注意、知识、学习、决策、交互等。
<B>'''技术手段:</B>'''深度学习(Deep Learning)+ 增强学习(Reinforcement Learning)
深度学习(Deep Learning) + 增强学习(Reinforcement/Unsupervised Learning)== 前沿应用进展 ==
=前沿应用进展== 自然语言理解 ===
==自然语言理解==自然语言理解(Natural Language Understanding),使用的技术称为自然语言处理(Natural Language Processing,简称NLP)。
自然语言理解(Natural Language Understanding),使用的技术称为自然语言处理(Natural Language Processing,简称NLP) ===智能问答===
整合语音识别ASR,计算机视觉CV和自然语言处理NLP的问答系统QA。
相关阅读:
Reasoning in vector space: An exploratory study of question answering, ICLR 2016.
=其它课程=相关课程:
[[实验室探究课-智能问答与智能系统]]