Friday, September 19, 2008

Automatic music emotions classification

There are 8 categories of music emotions.

1. Sad
2. Happy
3. Interesting

and there are 5 mores.

The following features are used to classify music emotions

1. Tempo
2. Pitches

Automatic music emotions classification

There are 8 categories of music emotions.

1. Sad
2. Happy
3. Interesting

and there are 5 mores.

The following features are used to classify music emotions

1. Tempo
2. Pitches

Thursday, September 18, 2008

Telecommunications evolution

1. Network transformation (e.g. time-based switches to packet-based switches).
2. Service providers transformation (e.g. voice and limited data service on public switched telephone network (PSTN) vs. voice and broadband-based voice and data on fixed and wireless networks).
3. Service transformation (e.g. voice over copper to voice over IP networks, TVs over IP networks, tele-video conferences).

Telecommunications evolution

1. Network transformation (e.g. time-based switches to packet-based switches).
2. Service providers transformation (e.g. voice and limited data service on public switched telephone network (PSTN) vs. voice and broadband-based voice and data on fixed and wireless networks).
3. Service transformation (e.g. voice over copper to voice over IP networks, TVs over IP networks, tele-video conferences).

Wednesday, September 17, 2008

Data representation

To represent a source of data in a compact form
1. Using the affine transform
2. Using the fractal transform
3. Using the Fourier transform
4. Using the wavelet transform

Data representation

To represent a source of data in a compact form
1. Using the affine transform
2. Using the fractal transform
3. Using the Fourier transform
4. Using the wavelet transform

Transform coding

1. Divide the sequence {x_n} into N blocks. Each block is mapped into a transformed sequence using a reversible mapping.
2. Quantizing the transform sequence. The quantization strategy depends on three main factors: the desired average bit rate, the statistics of the elements of transformed sequence and the effect of distortion in the transformed sequence on the reconstructed sequence.
3. Encode the quantized values using some binary encoding technique.

Transform coding

1. Divide the sequence {x_n} into N blocks. Each block is mapped into a transformed sequence using a reversible mapping.
2. Quantizing the transform sequence. The quantization strategy depends on three main factors: the desired average bit rate, the statistics of the elements of transformed sequence and the effect of distortion in the transformed sequence on the reconstructed sequence.
3. Encode the quantized values using some binary encoding technique.

Tuesday, September 16, 2008

Vector quantization

I. How is VQ designed?

An image is divided into different blocks (e.g. 4x4 pixels). We use a codebook of length 16, 32, 64, 256. Sending the codebook, the index of the codebook to the receiver for decompression. Initialization for the codebook consists of 2 different strategies

1. Choose one vector as the averaged vector of the training set. Then split the training set into two sets, four sets and eight sets according to the requirements of the length of the codebook (i.e. 16, 32, 64)
2. Combine two vectors into one if they contribute a smallest distortion, continue until we have the required codebook.

II. What is VQ used for?

-Speech coding
-Image compression in low bit rates

Vector quantization

I. How is VQ designed?

An image is divided into different blocks (e.g. 4x4 pixels). We use a codebook of length 16, 32, 64, 256. Sending the codebook, the index of the codebook to the receiver for decompression. Initialization for the codebook consists of 2 different strategies

1. Choose one vector as the averaged vector of the training set. Then split the training set into two sets, four sets and eight sets according to the requirements of the length of the codebook (i.e. 16, 32, 64)
2. Combine two vectors into one if they contribute a smallest distortion, continue until we have the required codebook.

II. What is VQ used for?

-Speech coding
-Image compression in low bit rates

Monday, September 15, 2008

New trends in designing mobile phones

1. Each mobile phone can become a wifi hot-spot which can be used for several users to connect to the Internet.
2. Locations-based services (e.g. movies, restaurants, petrol stations, games centers, shopping centers, etc.).
3. More features are added (e.g. HD video recorders, high resolution cameras, higher optical zooms, etc.).
4. Longer battery life.
5. Prettier appearance (e.g. different shapes and forms).

New trends in designing mobile phones

1. Each mobile phone can become a wifi hot-spot which can be used for several users to connect to the Internet.
2. Locations-based services (e.g. movies, restaurants, petrol stations, games centers, shopping centers, etc.).
3. More features are added (e.g. HD video recorders, high resolution cameras, higher optical zooms, etc.).
4. Longer battery life.
5. Prettier appearance (e.g. different shapes and forms).

Saturday, September 13, 2008

Mathematical spaces and other concepts

1. Metric spaces (e.g. scalars, vectors, sequences, functions)
2. Norm spaces
3. Vector spaces
4. Inner product spaces
5. Functions -> sequence of functions
6. Number -> sequence of numbers

Mathematical spaces and other concepts

1. Metric spaces (e.g. scalars, vectors, sequences, functions)
2. Norm spaces
3. Vector spaces
4. Inner product spaces
5. Functions -> sequence of functions
6. Number -> sequence of numbers

Mathematical spaces and other concepts

1. Metric spaces (e.g. scalars, vectors, sequences, functions)
2. Norm spaces
3. Vector spaces
4. Inner product spaces
5. Functions -> sequence of functions
6. Number -> sequence of numbers

Friday, September 12, 2008

Naive Bayes classifier

1. Continuous numerical attributes need to use Normal distribution to compute P(income=12000|Yes), etc. The set of values of the continuous numerical attribute is used to find out the mean and variance.

Naive Bayes classifier

1. Continuous numerical attributes need to use Normal distribution to compute P(income=12000|Yes), etc. The set of values of the continuous numerical attribute is used to find out the mean and variance.

K-means and k-modes

1. When compare two values of a nominal attribute if they are equal the result is zero otherwise the result will be 1/n (n is the number of values of this nominal attribute).

K-means and k-modes

1. When compare two values of a nominal attribute if they are equal the result is zero otherwise the result will be 1/n (n is the number of values of this nominal attribute).

Tuesday, September 9, 2008

Upcomming new technologies and other concepts

1. USB3 devices (10x faster than USB2: about 4.8Gbps)
2. Cloud computing
3. Machine learning involves how to write a program that can learn. It learns from examples and feedback. Machine learning techniques include neural networks. It usually deals with a small set of data. The data must be error-free and must be appropriate for the learning task. An example application of machine learning is the computer chess program.
4. Data mining concerns about how to extract knowledge from existing data. The output of data mining queries is the rules, classification or clusters.

Upcomming new technologies and other concepts

1. USB3 devices (10x faster than USB2: about 4.8Gbps)
2. Cloud computing
3. Machine learning involves how to write a program that can learn. It learns from examples and feedback. Machine learning techniques include neural networks. It usually deals with a small set of data. The data must be error-free and must be appropriate for the learning task. An example application of machine learning is the computer chess program.
4. Data mining concerns about how to extract knowledge from existing data. The output of data mining queries is the rules, classification or clusters.

Saturday, September 6, 2008

Trends in designing new laptops and how to buy them

1. Better battery time
2. Prettier appearance (e.g. color, material, etc.)
3. Lighter
4. Faster

Buying a new laptop needs to consider the followings

1. Battery time
2. Weight
3. Memory (e.g. max 4GB for current 32 bits OS)
4. CPU (e.g. duo core, core 2 duo, quad core, atom, etc)
5. OS (e.g. XP, Vista, Linux, etc.)
6. Brand (e.g. Toshiba, IBM, Sony, Acer, Fujitsu, Dell, HP, etc.)

Trends in designing new laptops and how to buy them

1. Better battery time
2. Prettier appearance (e.g. color, material, etc.)
3. Lighter
4. Faster

Buying a new laptop needs to consider the followings

1. Battery time
2. Weight
3. Memory (e.g. max 4GB for current 32 bits OS)
4. CPU (e.g. duo core, core 2 duo, quad core, atom, etc)
5. OS (e.g. XP, Vista, Linux, etc.)
6. Brand (e.g. Toshiba, IBM, Sony, Acer, Fujitsu, Dell, HP, etc.)

Friday, September 5, 2008

How to design an optimal code

1. Kraft's inequality theorem tells us that if the Kraft's condition is satisfied, there exists a prefix code C. Also, if the code C is the prefix code, the Kraft's condition is satisfied.

2. The average length of the code is always greater than the entropy of the source S. Therefore, the better code is when L(C) is closer to H(S).

H(S) <= L(C)

where the source S has the probability distribution as P(S)={p1, p2, p3,..,pn}
L(C)=Sum of {l1 x p1 + l2 x p2 + l3 x p3 + ...+ ln x pn}

How to design an optimal code

1. Kraft's inequality theorem tells us that if the Kraft's condition is satisfied, there exists a prefix code C. Also, if the code C is the prefix code, the Kraft's condition is satisfied.

2. The average length of the code is always greater than the entropy of the source S. Therefore, the better code is when L(C) is closer to H(S).

H(S) <= L(C)

where the source S has the probability distribution as P(S)={p1, p2, p3,..,pn}
L(C)=Sum of {l1 x p1 + l2 x p2 + l3 x p3 + ...+ ln x pn}

Thursday, September 4, 2008

How to improve our chance to achieve a career as an IT worker

The following helps us to improve our chance to work as an IT worker.

1. Good system designer
2. Good communicator
3. Good collaborator

How to improve our chance to achieve a career as an IT worker

The following helps us to improve our chance to work as an IT worker.

1. Good system designer
2. Good communicator
3. Good collaborator

What is computer's operating system?

1. Like a government to provide services for users (i.e. users' applications)
2. Like police to control traffic (e.g. I/O device controllers)
3. Like a facilitator between hardware machine code and application software
4. Provide networking stack for communications
5. Manage physical and virtual memory
6. Manage users
7. Schedule CPU(s) times for processes

What is computer's operating system?

1. Like a government to provide services for users (i.e. users' applications)
2. Like police to control traffic (e.g. I/O device controllers)
3. Like a facilitator between hardware machine code and application software
4. Provide networking stack for communications
5. Manage physical and virtual memory
6. Manage users
7. Schedule CPU(s) times for processes

Wednesday, September 3, 2008

Golden rules in chess

1. Attack the center
2. Look for danger
3. Move our fangs (i.e. knights or bishops) out early
4. Castle early

Golden rules in chess

1. Attack the center
2. Look for danger
3. Move our fangs (i.e. knights or bishops) out early
4. Castle early

Tuesday, September 2, 2008

Collaborative technologies

The following are some of current collaborative technologies

1. Instant messaging
2. Emails
3. Webcam
4. Voice over IP
5. Video conferencing
6. Blogs
7. Wikis
8. Videos

Collaborative technologies

The following are some of current collaborative technologies

1. Instant messaging
2. Emails
3. Webcam
4. Voice over IP
5. Video conferencing
6. Blogs
7. Wikis
8. Videos

Different ways to defend our attacked piece in chess

1. Move it out of danger
2. Protect it by another piece
3. Block it with a minor piece
4. Capture the attacking piece
5. Capture another piece that is more important than our attacked piece
6. Do a check or checkmate

Different ways to defend our attacked piece in chess

1. Move it out of danger
2. Protect it by another piece
3. Block it with a minor piece
4. Capture the attacking piece
5. Capture another piece that is more important than our attacked piece
6. Do a check or checkmate

Monday, September 1, 2008

Upcoming products

1. Thin TVs from Sony (about 2cm thickness).
2. Bluetooth TV (sending images from mobile phone to TV using Bluetooth connection).
3. DVD upscaling from SD to HD.
4. 3G IPhone problem between Wifi data and mobile network data.
5. TVs with Yahoo widgets (e.g. weather information, movies information, sport scores).

Upcoming products

1. Thin TVs from Sony (about 2cm thickness).
2. Bluetooth TV (sending images from mobile phone to TV using Bluetooth connection).
3. DVD upscaling from SD to HD.
4. 3G IPhone problem between Wifi data and mobile network data.
5. TVs with Yahoo widgets (e.g. weather information, movies information, sport scores).

Mounting USB drives in Windows Subsystem for Linux

Windows Subsystem for Linux can use (mount): SD card USB drives CD drives (CDFS) Network drives UNC paths Local storage / drives Drives form...