[Discussion: Selection and analysis for mid-z Quasars] 
chair: Akiyama
*Analysis (science issues)
- QSO/AGN luminosity function
- QSO-QSO clustering
- small scale QSO pairs (lensed QSOs)
- QSO-galaxy clustering
- galaxy distribution vs. QSOs
[Questions]
What these observable, regarding SMBH properties, 
What can be directly/indirectly derived ?
What can be directly/indirectly constrained 
[What we want to derive from HSC data]
Accretion rate, duty-cycle, Halo occupation distribution, 
Trigger mechanism, bias, galaxy ancestor/descendant analysis

[Comments]
1.galaxy-AGN connection
Important to examine host properties (mass, color, metallicity, etc) 
->should discuss with galaxy evolution people 

2. Lensed QSOs
- AGN variability and micro-lensing are interesting
- most lensed QSOs are faint end of AGN LF: Why ?
- binary QSOs : BH merging events ?
- the fraction of small separation pairs is very small
=> we need more information of close binary QSOs

3. QSO/AGN luminosity function
On this issue, it is important to compare with theoretical model.
(More discussion, tomorrow morning session)

*Selection method (Technical issues)
[Completeness of the stellarity selection for QSOs] 
We may miss "extended sources at high-z" due to the measurement errors.
We may consider the contamination of "low-z compact galaxy"
We need some calculations whether we can distinguish between point sources 
and extended sources (Matsuoka-san, Nagoya Univ.).

[Accuracy of photometric redshift]
Improvements are necessary for z < 3 AGNs: very hard!
Probably, the following two items are useful. 
(i) Multi-wavelength selection is important (next section in details)
(ii) SDSS sample with HSC faint galaxies at z=0.5-1 (Hashimoto-san)
#On this issue, we need short exposure avoiding the saturation.
#It is helpful to add the color of host galaxies (Tanaka-san).

[Technical issues] 
Feedback to Furusawa-san 
-Since AGNs are rare targets compared with galaxies and stars, 
only to write wonderful papers but also to understand the date quality 
and to improve software are also important. 
- How to detect QSO pairs
 --development algorithm 
 --Important to construct the catalog of HSC data: JVO-team can help this issue

-Exposure map, mask  (exptime, vignetting, survey-are)
Shirasaki-san, Komiya-san; Clustering analysis from JVO
They didn't consider mask pattern and exposure map 
because they use good data.

[1st science paper]
- Important to examine individual targets in detail
- collaboration with galaxy team: COSMOS, Spitzer, etc