towards recurrent neural network language models with linguistic and contextual features

... lines and j rows with binary values. The value zij = 1 (zij = 0) means that the word i influences (not) the word j. In figure 1 every link stands for zij = l. The models 1, 2 and 2 ~ and ... COMB), we determine TL WCs with the first method and afterwards we determine SL WCs with the sec- ond method. So we can use the well known iterative method to end up with WCs in different lan- ... Pietra, and R. L. Mercer. 1993. The mathematics of statistical machine transla- tion: Parameter estimation. Computational Linguistics, 19(2):263-311. P. Fung and D. Wu. 1995. Coerced markov models...

Ngày tải lên: 08/03/2014, 05:21

Ngày tải lên: 17/03/2014, 22:20

... Computational Linguistics, pages 1288–1297, Portland, Oregon, June 19-24, 2011. c 2011 Association for Computational Linguistics Enhancing Language Models in Statistical Machine Translation with Backward ... lan- guage models and extend their ability to capture richer contexts and long-distance dependencies. In particular, we integrate backward n-grams and mu- tual information (MI) triggers into language models in ... Demuynck, and Patrick Wambacq. 2002. Confidence scoring based on back- ward language models. In Proceedings of ICASSP, pages 221–224, Orlando, FL, April. Ahmad Emami, Kishore Papineni, and Jeffrey...

Ngày tải lên: 07/03/2014, 22:20

... cand[v] is undefined then 3: cand [v] ← ∅ 4: FIRE(e, 1, cand) foreach e ∈ IN (v) 5: buf [v] ← ∅ 6: while |D(v)| < j and |buf [v]| + |D(v)| < k and |cand[v]| > 0 do 7: item ← POP-MIN(cand[v]) 8: ... KBEST(v) 6: cand ← {e, 1 | e ∈ IN (v)} ⊲ for each incoming e 7: HEAPIFY(cand) ⊲ a priority queue of candidates 8: buf ← ∅ 9: while |cand| > 0 and |buf | < k do 10: item ← POP-MIN(cand) 11: ... node. 2.2 Adding a Language Model To integrate with a bigram language model, we can use the dynamic-programming algorithms of Och and Ney (2004) and Wu (1996) for phrase-based and SCFG-based systems,...

Ngày tải lên: 17/03/2014, 04:20

... left-hand side of the dashed line repre- sents an off-line network building process. The right-hand side represents a kana-kanji con- version process reinforced with a neural net- work handler. ... process reinforced with a neural net- work handler. The network is used by the neural network handler and word associations are done in parallel with kana-kanji conver- sion. The kana-kanji converter ... searches the dictio- nary for lexical and grammatical information and finally creates a list of possible homonym candidates. Then the neural network handler is requested for activation levels...

Ngày tải lên: 17/03/2014, 08:20

Ngày tải lên: 28/04/2014, 10:11

Ngày tải lên: 28/04/2014, 10:16

Ngày tải lên: 29/04/2014, 14:52

Ngày tải lên: 20/06/2014, 22:20

Ngày tải lên: 21/06/2014, 20:20

Ngày tải lên: 21/06/2014, 20:20

Ngày tải lên: 27/06/2014, 18:20

... 3 features of candidate text regions. Those features are used as the input of a neural network for training based on the back-propagation algorithm for neural networks. After the neural network ... the neural network 2.2 Neural Network In this subsection, text extraction from static image or video sequences is accomplished using the back-propagation (BP) algorithm on a neural network. ... network. The training of the neural network is based on the features we obtain from the DWT detail component sub-bands. As shown in Figure 6, the proposed neural network architecture is simpler...

Ngày tải lên: 05/11/2012, 14:51

... Together with the two basic spatial features, i.e., width and height, there are totally 14 features used for the neural network analysis. D. Contours classification using neural network We ... with fanciful design containing various graphical foreground and reverse contrast regions. The proposed method is to apply a neural network on canny edges with both spatial and relative features ... features for the neural network and confirm that the inter-relationship of neighboring characters is crucial in distinguishing most graphical objects from texts. Table 1 shows the recall and...

Ngày tải lên: 05/11/2012, 14:54

... planning and management of coastal systems and resources, taking into account traditional, cultural and historical perspectives and conflicting interests and uses; it is a continuous and evolutionary ... 1.6). Topology and Geology Topologically and geologically, the mainland of the study area can be separated into two regions: the lowlands in the Western part and highlands in the Eastern ... (Dowlatabadi, 1995; Hulme and Raper, 1995; Janssen and de Vries, 1998) and natural resources and environmental management (Stephens and Hess, 1999; Turner, 2000; De Kok and Wind, 2002). These...

Ngày tải lên: 06/11/2012, 10:35

Ngày tải lên: 05/09/2013, 14:58

Ngày tải lên: 12/09/2013, 16:20

... gọn như sau : Mạng nhiều lớp neuron CHƯƠNG 2 MÔ HÌNH MẠNG NEURAL NETWORKS Mô hình mạng Neural tổng quát có dạng như sau : Ngày nay mạng Neural có thể giải quyết nhiều vấn đề phức tạp đối với con ... định dạng của dữ liệu vào ảnh hưởng đến việc mô phỏng của mạng. Có hai loại mạng static network và dynamic network. Hai kiểu vector đầu vào cơ bản là kiểu xảy ra đồng thời (concurrently) và kiểu ... trong lớp a : vector ngõ ra của lớp neuron Hàm truyền Có rất nhiều hàm truyền áp dụng trong Neural Networks, trong đó ba hàm thường sử dụng nhất là Hard Limit, Linear, Log-Sigmoid. Tổng quát...

Ngày tải lên: 29/09/2013, 06:20