1 |
From sBoW to dCoT: Marginalized Encoders for Text Representation
|
|
|
|
In: http://www.cse.wustl.edu/~mchen/papers/msda_dm.pdf (2012)
|
|
BASE
|
|
Show details
|
|
2 |
From sBoW to dCoT: Marginalized Encoders for Text Representation
|
|
|
|
In: http://www.cse.wustl.edu/~mchen/papers/dCoT.pdf (2012)
|
|
BASE
|
|
Show details
|
|
3 |
Large margin training of continuous density hidden Markov models
|
|
|
|
In: http://www.cs.ucsd.edu/~saul/papers/lmb08_cdhmm.pdf (2009)
|
|
BASE
|
|
Show details
|
|
4 |
Comparison of large margin training to other discriminative methods for phonetic recognition by hidden Markov models
|
|
|
|
In: http://www.cs.berkeley.edu/~feisha/pubs/icassp2007.pdf (2007)
|
|
BASE
|
|
Show details
|
|
5 |
Comparison of large margin training to other discriminative methods for phonetic recognition by hidden markov models
|
|
|
|
In: http://www.cs.ucsd.edu/~saul/papers/icassp07_margin.pdf (2007)
|
|
BASE
|
|
Show details
|
|
6 |
Large margin hidden Markov models for automatic speech recognition
|
|
|
|
In: http://books.nips.cc/papers/files/nips19/NIPS2006_0143.pdf (2007)
|
|
BASE
|
|
Show details
|
|
7 |
Large margin hidden Markov models for automatic speech recognition
|
|
|
|
In: http://www.cs.berkeley.edu/~feisha/pubs/nips2006.pdf (2007)
|
|
BASE
|
|
Show details
|
|
8 |
Shallow Parsing with Conditional Random Fields
|
|
|
|
In: http://acl.ldc.upenn.edu/N/N03/N03-1028.pdf (2003)
|
|
BASE
|
|
Show details
|
|
9 |
Shallow parsing with conditional random fields
|
|
|
|
In: http://www.cis.upenn.edu/~pereira/papers/shallow.pdf (2003)
|
|
BASE
|
|
Show details
|
|
10 |
Shallow Parsing with Conditional Random Fields
|
|
|
|
In: http://www.cis.upenn.edu/~feisha/pubs/shallow03.pdf (2003)
|
|
BASE
|
|
Show details
|
|
11 |
Diverse Sequential Subset Selection for Supervised Video Summarization
|
|
|
|
In: http://www-scf.usc.edu/%7Eboqinggo/Paper/nips14_seqdpp.pdf
|
|
BASE
|
|
Show details
|
|
12 |
Supplementary Material Geodesic Flow Kernel for Unsupervised Domain Adaptation
|
|
|
|
In: http://www.cs.utexas.edu/~grauman/papers/subspace-supp-cvpr2012.pdf
|
|
Abstract:
Throughout the main text, many algorithmic details and empirical results were omitted and only discussed briefly so as to observe the limit on the number of pages. In this file, we expand the discussions in the main text and provide more details on • the derivation of our geodesic flow kernel (GFK) (Sec. A), explaining how eq.(5) and eq.(6) in the main text are derived (section 3.3). • how to compute the rank of domain (ROD) metric (Sec. B); the idea was only sketched in section 3.5 of the main text. • empirical studies of domain adaptation between 3 do-mains: Amazon, DSLR and Webcam (Sec. C). We had conducted two parallel empirical studies, one on the 3 domains and the other one on the 4 domains obtained from expanding the 3 with the dataset of Caltech-256. While both sets of empirical studies have reached the same findings that validate our methods, we chose to focus on domain adaptation among the 4 domains to demonstrate that our methods are robust to the additional diversity beyond the original 3. To be comprehensive, we report our results on those 3 domains as they provide a worthy reference point to contrast our work directly to published ones. • empirical studies of domain adaptation between Cal-tech, Amazon, Webcam and DSLR (Sec. D). In the main text (section 4), we reported only 8 of 12 pos-sible pairs of source and target domains. This Suppl. reports the remaining 4 pairs. • characterizing the datasets of PASCAL, ImageNet, and Caltech-101 with our ROD metric (Sec. E). The met-ric corroborates our empirical findings on the cross-dataset generalization performances of these 3 do-mains (section 4.6, and especially Table 3). A. The derivation of the geodesic flow kernel Let ΩT denote the following matrix ΩT = [PS RS]
|
|
URL: http://www.cs.utexas.edu/~grauman/papers/subspace-supp-cvpr2012.pdf http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.646.9531
|
|
BASE
|
|
Hide details
|
|
|
|