Depth Prediction Without the Sensors: Leveraging Structure for Unsupervised L...harmonylab
紹介論文
Depth Prediction Without the Sensors: Leveraging Structure for Unsupervised Learning from Monocular Videos
出典: Vincent Casser, Soeren Pirk Reza, Mahjourian, Anelia Angelova : Depth Prediction Without the Sensors: Leveraging Structure for Unsupervised Learning from Monocular Videos, the AAAI Conference on Artificial Intelligence, Vol. 33, pp. 8001-8008 (2019)
概要: カメラ映像による深度予測は、屋内及び屋外のロボットナビゲーションにとって必要なタスクです。本研究では、教師なし学習を用いて映像の深度予測とカメラのエゴモーション(自身の動き)の学習に取り組んでいます。先行研究で確立されたベースラインのモデルに、移動する個々の物体のモデル化と、オンラインでのモデルの調整を行う手法を取り入れています。結果として、物体の動きを多く含むシーンでの予測結果を大幅に向上させています。
Depth Prediction Without the Sensors: Leveraging Structure for Unsupervised L...harmonylab
紹介論文
Depth Prediction Without the Sensors: Leveraging Structure for Unsupervised Learning from Monocular Videos
出典: Vincent Casser, Soeren Pirk Reza, Mahjourian, Anelia Angelova : Depth Prediction Without the Sensors: Leveraging Structure for Unsupervised Learning from Monocular Videos, the AAAI Conference on Artificial Intelligence, Vol. 33, pp. 8001-8008 (2019)
概要: カメラ映像による深度予測は、屋内及び屋外のロボットナビゲーションにとって必要なタスクです。本研究では、教師なし学習を用いて映像の深度予測とカメラのエゴモーション(自身の動き)の学習に取り組んでいます。先行研究で確立されたベースラインのモデルに、移動する個々の物体のモデル化と、オンラインでのモデルの調整を行う手法を取り入れています。結果として、物体の動きを多く含むシーンでの予測結果を大幅に向上させています。
Statstical Genetics Summer School 2023
http://www.sg.med.osaka-u.ac.jp/school_2023.html
Aug 25-27th 2023, Osaka University, The University of Tokyo, RIKENm, Japan
Changsheng Zhang, Bo Tang, Qian Wang and Luhua Lai.
Discovery of binding proteins for a protein target using protein-protein docking-based virtual screening.
Proteins, 2014 (early access on May 26)
ライフサイエンスデータベースの現状 〜データベース統合化のための技術的・政治的側面〜
Japan Museum Bioinformatics (Museomics) Working Group 第2回会合@東工大・緑が丘キャンパス
https://sites.google.com/site/museumbioinfo/meetings/201410xx
#museomejp
Statstical Genetics Summer School 2023
http://www.sg.med.osaka-u.ac.jp/school_2023.html
Aug 25-27th 2023, Osaka University, The University of Tokyo, RIKENm, Japan
Changsheng Zhang, Bo Tang, Qian Wang and Luhua Lai.
Discovery of binding proteins for a protein target using protein-protein docking-based virtual screening.
Proteins, 2014 (early access on May 26)
ライフサイエンスデータベースの現状 〜データベース統合化のための技術的・政治的側面〜
Japan Museum Bioinformatics (Museomics) Working Group 第2回会合@東工大・緑が丘キャンパス
https://sites.google.com/site/museumbioinfo/meetings/201410xx
#museomejp
Predicting protein–protein interactions based only on sequences information
Juwen Shen, Jian Zhang, Xiaomin Luo, Weiliang Zhu, Kunqian Yu, Kaixian Chen, Yixue Li and Hualiang Jiang
Proc Natl Acad Sci USA, 2007, 104(11), 4337-4341.
3. ISMB/ECCB
2015
• Joint
conference
– 23rd
annual
mee=ng
of
Intelligent
Systems
for
Molecular
Biology
(ISMB)
– 14th
European
Conference
on
Computa=onal
Biology
(ECCB)
• 開催地: ダブリン(アイルランド)
• 日程:
7月10〜14日
• プロシーディング:
Bioinforma=cs誌の特別号
4. ISMB/ECCB
2015
• 採択率:
42
/
241
≒
17.4%
and students in the field. The 42 papers in this volume were selected
from 241 original submissions divided into 13 research areas, col-
lectively led by 25 Area Chairs. For each area, the Area Chairs se-
lected an expert program committee for their subdiscipline and
oversaw the reviewing process for that area. By design, the Area
Chairs included a mix of experienced individuals reappointed from
previous years and experts newly recruited to ensure broad tech-
nical expertise and to promote inclusivity of various elements of the
research community. In total, the review process involved the 25
Area Chairs, 378 program committee members, and an additional
27 papers that were resubmitted, 15 were judged to have addressed
the concerns of the reviewers and were accepted for the conference
proceedings, resulting in a total of 42 acceptances and an overall
acceptance rate of 42/241 ¼ 17.4%. We believe that this two-tier
system, which is more reflective of typical multi-round journal re-
view procedures, provided a means of ensuring that only the high-
est quality original work was accepted within the tight timing
constraints imposed by the conference scheduling. We thank all
authors for submitting their work. These proceedings would sim-
ply not be possible without the scientific ingenuity of the
Table 1. ISMB/ECCB 2015 review summary by area.
Topic area Chairs Submissions Accepted
round 1
Invited for
round 2
Accepted
in round 2
Approved for
proceedings
Applied Bioinformatics Thomas Lengauer and Christophe
Dessimoz
30 1 6 3 4
Bioimaging and Data Visualization Robert Murphy 12 1 2 1 2
Databases, Ontologies and Text Mining Hagit Shatkay and Helen Parkinson 11 1 1 1 2
Disease Models and Epidemiology Simon Kasif and Alice McHardy 21 3 3 3 6
Evolution and Comparative Genomics Bernard Moret and Louxin Zhang 12 2 0 0 2
Gene Regulation and Transcriptomics Uwe Ohler and Zohar Yakhini 30 2 4 2 4
Mass Spectrometry and Proteomics Olga Vitek and Knut Reinert 11 2 0 0 2
Metabolic Networks Bonnie Berger and Hidde de Jong 5 2 0 0 2
Population Genomics Russell Schwartz and Jennifer
Listgarten
22 3 2 1 4
Protein Interactions and Molecular
Networks
Natasa Przulj and Igor Jurisica 29 2 5 3 5
Protein Structure and Function Torsten Schwede and Anna
Tramontano
22 3 2 1 4
RNA Bioinformatics Jerome Waldispuhl and Hanah
Margalit
6 0 1 0 0
Sequence Analysis Michael Brudno and Siu-Ming Yiu 30 5 3 0 5
241 27 29 15 42
7. 慶應義塾大学理工学部
佐藤健吾
satoken@bio.keio.ac.jp
Misassembly detection using paired-end
sequence reads and optical mapping data
Martin D. Muggli1,
*, Simon J. Puglisi2
, Roy Ronen3
and
Christina Boucher1
1
Department of Computer Science, Colorado State University, Fort Collins, CO 80526, USA, 2
Department of
Computer Science, University of Helsinki, Finland and 3
Bioinformatics Graduate Program, University of California,
San Diego, La Jolla, CA 92093, USA
*To whom correspondence should be addressed.
Abstract
Motivation: A crucial problem in genome assembly is the discovery and correction of misassembly
errors in draft genomes. We develop a method called MISSEQUEL that enhances the quality of draft
genomes by identifying misassembly errors and their breakpoints using paired-end sequence
reads and optical mapping data. Our method also fulfills the critical need for open source computa-
tional methods for analyzing optical mapping data. We apply our method to various assemblies of
the loblolly pine, Francisella tularensis, rice and budgerigar genomes. We generated and used
stimulated optical mapping data for loblolly pine and F.tularensis and used real optical mapping
data for rice and budgerigar.
Results: Our results demonstrate that we detect more than 54% of extensively misassembled con-
tigs and more than 60% of locally misassembled contigs in assemblies of F.tularensis and between
31% and 100% of extensively misassembled contigs and between 57% and 73% of locally misas-
Bioinformatics, 31, 2015, i80–i88
doi: 10.1093/bioinformatics/btv262
ISMB/ECCB 2015
ISMB/ECCB
2015読み会@東大
10. Recruitment
of
reads
• ペアエンドリードをコンティグに貼り付ける。
• 張り付いた順番、向き、カバレッジから、ミス
アセンブリの候補を検出する。
Correct assembly
A R CR A R CR
Inversion
A R CR
mate-pair 1 mate-pair 2 mate-pair 3
mate-pair 1 mate-pair 2 mate-pair 3
mate-pair 1 mate-pair 2 mate-pair 3
Rearrangment
A R CR
v
Correct assembly (read depth)
A R C
v
Collapsed repeat Expanded repeat:
A R CRR
v
(a) (b) (c)
(d) (e) (f)
Fig. 1. An illustration about the systematic alterations that occur with rearrangements, inversions, collapsed repeats and expanded repeats. (a) Proper read align-
ment where mate-pair reads have the correct orientation and distance from each other. A rearrangement or inversion will present itself by the orientation of the
reads being incorrect and/or the distance of the mate-pairs being significantly smaller or significantly larger than the expected insert size. This is shown in (b) and
i82 M.D.Muggli et al.
11. Red-‐black
posi=onal
de
Bruijn
graph
• コンティグ上の位置情報を持ったde
Bruijn
graph
• カバレッジが近傍の平均と比べて極端に外れている
場合、ペアエンドの整合性が取れない場合⇒赤、そ
うでない場合⇒黒
• 赤が50個以上続いたらミスアセンブリ候補
Misassembly detection using paired-end sequence reads and optical mapping data i83