Search results for "artificial intelligence"
showing 10 items of 6122 documents
Discovering the Senses of an Ambiguous Word by Clustering its Local Contexts
2005
As has been shown recently, it is possible to automatically discover the senses of an ambiguous word by statistically analyzing its contextual behavior in a large text corpus. However, this kind of research is still at an early stage. The results need to be improved and there is considerable disagreement on methodological issues. For example, although most researchers use clustering approaches for word sense induction, it is not clear what statistical features the clustering should be based on. Whereas so far most researchers cluster global co-occurrence vectors that reflect the overall behavior of a word in a corpus, in this paper we argue that it is more appropriate to use local context v…
Weights Space Exploration Using Genetic Algorithms for Meta-classifier in Text Document Classification
2012
A Controllable Text Simplification System for the Italian Language
2021
Text simplification is a non-trivial task that aims at reducing the linguistic complexity of written texts. Researchers have studied the problem by proposing new methodologies for addressing the English language, but other languages, like the Italian one, are almost unexplored. In this paper, we give a contribution to the enhancement of the Automated Text Simplification research by presenting a deep learning-based system, inspired by a state of the art system for the English language, capable of simplifying Italian texts. The system has been trained and tested by leveraging the Italian version of Newsela; it has shown promising results by achieving a SARI value of 30.17.
Cost-driven framework for progressive compression of textured meshes
2019
International audience; Recent advances in digitization of geometry and radiometry generate in routine massive amounts of surface meshes with texture or color attributes. This large amount of data can be compressed using a progressive approach which provides at decoding low complexity levels of details (LoDs) that are continuously refined until retrieving the original model. The goal of such a progressive mesh compression algorithm is to improve the overall quality of the transmission for the user, by optimizing the rate-distortion trade-off. In this paper, we introduce a novel meaningful measure for the cost of a progressive transmission of a textured mesh by observing that the rate-distor…
Copy-move Forgery Detection via Texture Description
2010
Copy-move forgery is one of the most common type of tampering in digital images. Copy-moves are parts of the image that are copied and pasted onto another part of the same image. Detection methods in general use block-matching methods, which first divide the image into overlapping blocks and then extract features from each block, assuming similar blocks will yield similar features. In this paper we present a block-based approach which exploits texture as feature to be extracted from blocks. Our goal is to study if texture is well suited for the specific application, and to compare performance of several texture descriptors. Tests have been made on both uncompressed and JPEG compressed image…
New techniques for visualization of losses due to image compression in grayscale medical still images
2003
To evaluate the visual influence of irreversible compression on medical images, changes of the images have to be visualized. The authors have explored alternative techniques to be used instead of the usual side-by-side comparison, where the information contained in both images is perceived in a single image, preserving the context between compression errors and image structures. Thus fast and easy comparison can be done. These techniques make use of the human ability to perceive information also in the dimensions of color, space, and time. A study was performed with JPEG-compressed coronary angiographic images. Changes in the resulting images for six compression factors from 7 to 30 were sc…
Sähköä ja alkemiaa: Tekoälydiskurssit Yleisradion verkkoartikkeleissa
2021
Tässä artikkelissa tarkastelemme sitä, millaisena ja miten tekoäly esitetään suomalaisessa julkisessa keskustelussa, ja ketkä tekoälystä suurelle yleisölle kertovat. Aineistona olemme käyttäneet Yleisradion verkkosivujen tekoälyä käsitteleviä artikkeleja. Tulosten perusteella tekoälystä pääsevät kertomaan useimmin talouden ja teollisuuden aloilla työskentelevät miehet. Aineistossa esiintyneet tekoälytulevaisuuskuvaukset pitäytyvät pitkälti nykyisen länsikeskeisen kapitalistisen maailmankuvan sisällä. Toisin sanoen, ne eivät haastaneet tai ylittäneet vallitsevaa status quoa, vaan näkivät tulevaisuuden pikemminkin lineaarisena kehityksenä nykytilasta. Aiemman tutkimuksen perusteella yleinen k…
Graph Clustering with Local Density-Cut
2018
In this paper, we introduce a new graph clustering algorithm, called Dcut. The basic idea is to envision the graph clustering as a local density-cut problem. To identify meaningful communities in a graph, a density-connected tree is first constructed in a local fashion. Building upon the local intuitive density-connected tree, Dcut allows partitioning a graph into multiple densely tight-knit clusters effectively and efficiently. We have demonstrated that our method has several attractive benefits: (a) Dcut provides an intuitive criterion to evaluate the goodness of a graph clustering in a more precise way; (b) Building upon the density-connected tree, Dcut allows identifying high-quality cl…
On the impact of forgetting on learning machines
1995
People tend not to have perfect memories when it comes to learning, or to anything else for that matter. Most formal studies of learning, however, assume a perfect memory. Some approaches have restricted the number of items that could be retained. We introduce a complexity theoretic accounting of memory utilization by learning machines. In our new model, memory is measured in bits as a function of the size of the input. There is a hierarchy of learnability based on increasing memory allotment. The lower bound results are proved using an unusual combination of pumping and mutual recursion theorem arguments. For technical reasons, it was necessary to consider two types of memory : long and sh…
A SOM/ARSOM Hierarchy for the Description of Dynamic Scenes
2001
A neural architecture is presented, aimed to describe the dynamic evolution of complex structures inside a video sequence. The proposed system is arranged as a tree of self-organizing maps. Leaf nodes are implemented by ARSOM networks as a way to code dynamic inputs, while classical SOM's are used to implement the upper levels of the hierarchy. Depending on the application domain, inputs are made by suitable low level features extracted frame by frame of the sequence. Theoretical foundations of the architecture are reported along with a detailed outline of its structure, and encouraging experimental results.