Animal Learning Vs Machine Learning

Learning, like intelligence covers such a broad range of processes that is difficult to define. Machine learning has much similarity with animal learning. There are several parallels between machine and animal learning. Many techniques in machine learning are derived from efforts to make more theories of animal and machine learning through computational model.

A machine learns whenever it changes its structure, program, or data in such a manner that its future performance improves. The changes, like addition of a record in a database seems quite easy, but when a performance of speech recognition machine improves after hearing sample of person’s speech, then we can feel justified that the machine has learned. [5]

Machine learning usually refers to the changes in systems that perform tasks associated with Artificial Intelligence. These tasks include recognition, diagnosis, robot control, prediction etc.

[1] Machine Learning is a field of computer science that deals with construction of systems that can learn from data, rather than following programmed instructions. In Computer Science it has strong ties with “Artificial Intelligence”. Machine learning focuses on prediction based on properties learned from data. [1]

Unsupervised learning is a type of machine learning where we are given outputs with no inputs. Unsupervised learning can also be called learning without a teacher. The goal here is to discover hidden patterns from data, which is called as “knowledge discovery.” An example of unsupervised learning is “topic modeling” where a list of documents is given and the task is to find out documents with similarity. It is more widely applicable than supervised learning. It does not require a human expert to manually enter the data. In contrast with the supervised learning or Reinforcement learning, there are no explicit target outputs; rather the unsupervised learner brings to bear prior biases as to what aspects of the structure of the input should be captured in the

For example spam filtering is an example of supervised learning. The most common are document classification and e-mail spam filtering.

Other examples include image classification, handwriting recognition, face detection and recognition.

Image classification is much harder problem because classifying an image directly is very difficult. Here human has not pre-processed the data. We might want to classify the image. Is it an indoor or a outdoor scene? Is it a vertical or horizontal photo? What does this image contain?

A much harder problem is to classify an image or detect something in a image or some object in a image. A special case in this problem is face detection. One approach to solve this problem is to divide the image into small patches at different locations and classify each patch whether it contain face like texture or not. This system then returns the locations where probability of being face is high. These face detection systems are now being used in modern camera, smart phones