2 years ago
The results of scientific researches received in recent years in problems of voice recognition , machine translation , determination of a shade of the sentence  and word classes  showed prospects of neural network algorithms of deep learning in comparison with classical methods of natural languag processing (natural language processing). However, in the field of question-answer and dialogue systems still there are many unresolved tasks [5, 6]. In this article the overview of results of application of modern algorithms for problems of processing and a natural language understanding is given. The overview contains the description of several different approaches and does not apply for completeness of researches.
Human: how many legs does a cat have? Machine: four, i think. Human: What do you think about messi? Machine: he ’s a great player. Human: where are you now? Machine: i ’m in the middle of nowhere.
2 years ago. In comparison with native JS on the DOM elements, implementation of animation algorithms on Canvas is usually more productive many times over. It is the known fact (but with features for small number of particles as it will become clear later), and it can find implementation so everything disturbing traditional under NG, but driven by rational users of "the falling snow". That it was not enough loading, in recent years it is considered a good form "to start" snow on the website scarcely noticeable, with the minimum quantity of snowflakes (5-15). Here both the effect is, and honor loads of the processor any.
Therefore, so far to NG some more days still winter, I suggest to arrange a hakatonchik on implementation of the best algorithms on a kanvasa and their analogs on DOM, having taken as a basis generally ancient native algorithms which as at most turned back in jQuery plug-in that it was convenient to connect. The most part of these algorithms does not proportion load of the processor or is made inefficiently therefore even at small number of snowflakes load the processor for 100%. Here an example of a review article where more than 10 implementations are considered, not everything, meeting in the nature. In addition, we will consider several electees to receive a backlog on development of algorithm and implementation it with good efficiency (5-6 more options will turn out). On this basis it is possible to construct completion.
In this post it will be a question of my participation in the tender Ludum Dare 34 which was about three weeks ago.
As a result the puzzle under the name Growing Sakura which gameplay it is possible to see on a gif image turned out (be not frightened, it weighs only 300 KB):
Briefly about rules of the game: initially we have a hexagonal field and several root buds (or one, as on a gif image above). From it it is possible to start up 3 branches (by two methods — clicking left or the right mouse button). On a branch the left click of a mouse can be made of each bud a Y-branching, and right — it is simple to continue a branch further (I-branching). If the branch cannot grow in any direction (the corresponding cage is occupied or in the necessary direction there is no cage) — that the branch does not grow. According to the last condition it is necessary to select correctly poryadkok "deployment" of branches. As a result the tree (or several trees) it that between two adjacent branches there are no acute angles will turn out. The game purpose — to cover all cages of a game field.
Without looking under kat try to think seconds 10 and to estimate how difficult can be this game.
2 years ago
Some time ago I was interested in a problem of sex determination of the person of his Full Name. At that moment I worked in the field of medical insurance where this problem was really actual – expenses on one insured, so and rates on which people were accepted on insurance, depending on a sex of the client, could differ several times. The most part of agreements – corporate, insured are the employer's employees.
We never saw the majority of them in eyes, everything that we had – lists insured where a floor was sometimes specified (with a large number of errors), but is more often – is not specified in general. The majority of the companies have the specifics of work and professional traditions owing to what, in their collectives same-gender people prevail. Even the small error could make potentially profitable agreement unprofitable (or on the contrary, but on it, on strange combination of circumstances, it happened to our clients much less often). In general, at the volume of a portfolio of agreements in several billions, and characteristic quantity of errors about a percent, the price of correct sex determination on the Full Name was around several tens millions.
2 years ago
Compressor units in the industry are used in many technology operations. Compressed air is received different types of compressor units. From rotor type, to vortex turbomachines. Centrifugal K-250 compressor units have a wide circulation in the industry. But all types of compressors have a critical duty of work – a surge.
2 years ago
I suggest to work out in a circle of great ladies-programmers a traditional New Year's task about 2016. It is necessary to place signs and brackets that any number from 1 to 100 turned out. For example
2 years ago
Being engaged in applications programming under Android OS there are interesting ideas which want to be tried, or there is some set of theoretical knowledge and they want to be put into practice, from set of these factors and there was an idea of the described project.
There are many articles about text recognition, about computer sight and about separate algorithms of recognition. In the same publication attempt of implementation of the task connected with finding of a key word on the image of the text that is able to afford to find, for example, the necessary place for reading any text in DjVu without recognition of the text is shown.
Example of implementation is presented in the form by application Android, and the source image is the text screenshot, with the entered key word, different algorithms of processing and image understanding are applied to a solution of a task.
So, we held the Festival of the new technologies Given at an exhibition here.
And we tell this first action from a series in which we bring together experts from different areas of business, science and public administration about analytics of data.
Storage and data analysis which were a prerogative of a narrow circle of the companies and people now begin to affect life practically all. For this reason we also began this series of actions where we tell wide audience about data and their analytics.
This article is devoted to surprising features in the world of chaos. I will try to tell about how to bridle such strange and difficult thing as chaotic process and to learn to create own elementary generators of chaos. Together with you we will pass a way from the dry theory before fine visualization of chaotic processes in space. In particular, on the example of the known chaotic attractors, I will show how to create dynamic systems and to use them in the tasks connected with the programmable logic integrated circuits (PLIC).
On December 26 our FlyElephant command will take part in Hub AI&BigData meetings; meetup devoted to big data and artificial intelligence. Action will take place in Odessa and will begin at 11:00. For all who will not be able to come online broadcasting will be organized.