Báo cáo khoa học: "Improving Search Results Quality by Customizing Summary Lengths" potx

... appropriate search result summary lengths, and that perceptions of search result quality can be affected by varying these result lengths. These findings have important impli- cations for search results ... USA, June 2008. c 2008 Association for Computational Linguistics Improving Search Results Quality by Customizing Summary Lengths Michael Kaisser University of Edin...

Ngày tải lên: 23/03/2014, 17:20

... modified by Haitao Mi and the English parser (Charniak and Johnson, 2005) modified by Liang Huang to produce en- tire parse forests. Then, we ran the Python scripts (Huang, 2008) provided by Liang ... for- est is first used by Huang and Chiang (2007) to characterize the search space of decoding with lan- guage models. The first direct use of packed for- est is proposed by Mi et al. (2...

Ngày tải lên: 23/03/2014, 16:21

... packages The learning process is also sustained by in- teractive elements, such as the possibility of changing parameters for the LSA algorithm and visualizing the results, or as the inte- grated programs ... algorithm. Initially thought for being used mostly by students from linguistics (or linguists) - due to the mathe- matical algorithms -, the tool can be exploited by anybody who w...

Ngày tải lên: 17/03/2014, 02:20

... embeddings that better capture the se- mantics of words by incorporating both local and global document context, and 2) accounts for homonymy and polysemy by learning mul- tiple embeddings per word. We ... syntac- tic information of words. These representations can be used to induce similarity measures by computing distances between the vectors, leading to many use- ful applications, su...

Ngày tải lên: 19/02/2014, 19:20

... resource. By interpolating CM1 and CM2, i.e. CM-3, the error rate of multi-word alignment results is further reduced. Figure 2 shows an example of word alignment results generated by the baseline ... translation quality improvement. The results are shown in Table 4. From the results of Table 4, it can be seen that the systems using the improved bi-directional alignments ach...

Ngày tải lên: 20/02/2014, 04:20

... results. Line 2 is the AI performance when gold candidate boundaries and word features are used; Line 3 is the perfor- mance with additional syntactic features. Line 4 shows the performance by ... features (denoted by †). Namely, word features play a more important role in SRC than in AI. Though the other eight features are based on full parsing, four of them (denoted by ‡) use the hea...

Ngày tải lên: 20/02/2014, 04:20

... an iteration no more results in a sig- nificant increase in the posterior probability. The search algorithm of Morfessor is a greedy algorithm where the costs of the next search points 33 Word-based ... Lagus, 2007). are affected by the decision in the current step. This leads to a sequential search and does not lend itself to parallelization. We propose a slightly modified search p...

Ngày tải lên: 20/02/2014, 04:20

... USA {ozaidan,ccb}@cs.jhu.edu Abstract Naively collecting translations by crowd- sourcing the task to non-professional trans- lators yields disfluent, low -quality results if no quality control is exercised. We demon- strate ... of professional translations provided by the LDC to non-professional translations created on Mechanical Turk. get high quality translations in aggregate by...

Ngày tải lên: 20/02/2014, 04:20

... provide computable information from narra- tive text and enable improved data quality and de- cision-making. Many NLP researchers working with clinical text (i.e. documents in the electronic health ... During the Fourth i2b2/VA Challenge, the asser- tion classification task was tackled by participating researchers. The best performing system (Berry de Bruijn et al., 2011) reached a m...

Ngày tải lên: 20/02/2014, 05:20

... that im- proves the quality of ASR transcripts for lectures. WER is reduced by 10% to 14%, with an average reduction of 12.9%, relative to initial values. This is achieved by making use of manual ... lecture transcription. This is in part caused by the mis- match between the language used in a lecture and the predictive language models employed by most ASR systems. Most ASR systems...

Ngày tải lên: 20/02/2014, 07:20