2,023 research outputs found
Short text authorship attribution via sequence kernels, Markov chains and author unmasking: An investigation
No full textWe present an investigation of recently proposed character and word sequence kernels for the task of authorship attribution based on relatively short texts. Performance is compared with two corresponding probabilistic approaches based on Markov chains. Several configurations of the sequence kernels are studied on a relatively large dataset (50 authors), where each author covered several topics. Utilising Moffat smoothing, the two probabilistic approaches obtain similar performance, which in turn is comparable to that of character sequence kernels and is better than that of word sequence kernels. The results further suggest that when using a realistic setup that takes into account the case of texts which are not written by any hypothesised authors, the amount of training material has more influence on discrimination performance than the amount of test material. Moreover, we show that the recently proposed author unmasking approach is less useful when dealing with short texts
Short Text Authorship Attribution via Sequence Kernels, Markov Chains and Author Unmasking: An Investigation
No full textWe present an investigation of recently proposed character and word sequence kernels for the task of authorship attribution based on relatively short texts. Performance is compared with two corresponding probabilistic approaches based on Markov chains. Several configurations of the sequence kernels are studied on a relatively large dataset (50 authors), where each author covered several topics. Utilising Moffat smoothing, the two probabilistic approaches obtain similar performance, which in turn is comparable to that of character sequence kernels and is better than that of word sequence kernels. The results further suggest that when using a realistic setup that takes into account the case of texts which are not written by any hypothesised authors, the amount of training material has more influence on discrimination performance than the amount of test material. Moreover, we show that the recently proposed author unmasking approach is less useful when dealing with short texts.
Short text authorship attribution via sequence kernels, Markov chains and author unmasking: An investigation
No full textWe present an investigation of recently proposed character and word sequence kernels for the task of authorship attribution based on relatively short texts. Performance is compared with two corresponding probabilistic approaches based on Markov chains. Several configurations of the sequence kernels are studied on a relatively large dataset (50 authors), where each author covered several topics. Utilising Moffat smoothing, the two probabilistic approaches obtain similar performance, which in turn is comparable to that of character sequence kernels and is better than that of word sequence kernels. The results further suggest that when using a realistic setup that takes into account the case of texts which are not written by any hypothesised authors, the amount of training material has more influence on discrimination performance than the amount of test material. Moreover, we show that the recently proposed author unmasking approach is less useful when dealing with short texts
Optimization of Weights in a Multiple Classifier Handwritten Word Recognition System Using a Genetic Algorithm
No full textAutomatic handwritten text recognition by computer has a number of interesting applications. However, due to a great variety of individual writing styles, the problem is very difficult and far from being solved. Recently, a number of classifier creation methods, known as ensemble methods, have been proposed in the field of machine learning. They have shown improved recognition performance over single classifiers. For the combination of these classifiers many methods have been proposed in the literature. In this paper we describe a weighted voting scheme where the weights are obtained by a genetic algorithm
The relation between religion and nationalism in early Zionist thought : a study in the prehistory of the State of Israel. --
No full textNET Models Meeting 2024 white paper: the current state of neuroendocrine tumour research models and our future aspirations
No full textCurrent models for the study of neuroendocrine tumours (NETs) are severely limited. While in vitro (e.g. cell lines), ex vivo (e.g. organoids) and in vivo (e.g. mice) models all exist, each has limitations. To address these limitations and collectively identify strategies to move the NET models field forward, we held an inaugural NET models meeting, hosted by our founding group: Dr Lines (Oxford), Prof. Quelle (Iowa), Dr Dayton (Barcelona), Dr Ear (Iowa), Dr Marinoni (Bern) and Dr Guenter (Alabama). This two-day meeting in Oxford (UK) was organised and supported by Bioscientifica Ltd and was solely dedicated to the discussion of NET models. The meeting was attended by ∼30 international researchers (from the UK, EU, Israel, USA and Canada). Plenary talks were given by Prof. Thakker, who summarised NET research over the past few decades, and Dr Schrader, who described the process and pitfalls of generating new cell lines. Eight researchers also presented their work on topics ranging from human cell 3D bioprinting to zebrafish models and included novel ideas and improvements on current concepts. This was followed by an interactive workshop, where discussion topics included a summary of currently available NET models, limitations of these models, barriers to developing new models, and how we can address these issues going forward. This white paper summarises the key points raised in these discussions and the future aspirations of the NET Models Consortium. The next meeting will take place in Oxford (UK) in 2025; contact [email protected] for more information.Depto. de Genética, Fisiología y MicrobiologíaFac. de Ciencias BiológicasTRUEpu
Short text authorship attribution via sequence kernels, Markov chains and author unmasking
No full textOn authorship attribution via Markov chains and sequence kernels
No full textWe investigate the use of recently proposed character and word sequence kernels for the task of authorship attribution and compare their performance with two probabilistic approaches based on Markov chains of characters and words. Several configurations of the sequence kernels are studied using a relatively large dataset, where each author covered several topics. Utilising Moffat smoothing, the two probabilistic approaches obtain similar performance, which in turn is comparable to that of character sequence kernels and is better than that of word sequence kernels. The results further suggest that when using a realistic setup that takes into account the case of texts which are not written by any hypothesised authors, about 5000 reference words are required to obtain good discrimination performance
Gary Schild Collection circa 1919-1999
No full textElectoral list for Jewish community of Magdeburg (1919); Schild family trees from 1738 to the 20th century; list of Jews from Geseke (before 1933); Information about the persons on a group photo on the cover of the book: Meyer, Beate: Jüdische Mischlinge: Rassenpolitik und Verfolgungserfahrung 1933-1945. Hamburg: Doelling und Gallitz, 1999. All items are photocopies.The following individuals are named in this collection:Fritz Blohm, Wolfgang Hecht, Hans-Peter Islar, Karl-Heinz Johrns, Heinz Loehnberg, Inge Meyer, Ursula Meyer, Ursula Pein, Guenter Schild, Herbert Simon, Gerhard WundermacherGary SchildProcessed for digitization byJewish communitiesHessedigitize
Sub-100 fs single-walled carbon nanotube saturable absorber mode-locked Yb-laser operation near 1 mu m
No full textTransmission-and reflection-type single-walled carbon nanotube saturable absorbers (SWCNT-SAs) were designed and fabricated for passive mode-locking of bulk lasers in the 1 mu m spectral range. Mode-locked laser operation based on a diffusion-bonded Yb:KYW/KYW crystal was demonstrated, and pulses as short as 83 fs and 140 fs were achieved applying reflection-type and transmission-type SWCNT-SA, respectively. The nonlinear parameters of the absorbers were measured to be in close vicinity to those of a semiconductor saturable absorber mirror for the same wavelength range. Mode-locking performance with SWCNT-SAs and the SESAM was compared utilizing the same cavity, with the SESAM resulting in only slightly shorter pulses of 66 fs duration. The nearly identical performance indicates that well-optimized SWCNT-SAs can substitute SESAMs even in the 1 mu m region. (C) 2009 Optical Society of Ameri
- …
