Xiangnan He
@South_River1
Followers
233
Following
26
Statuses
19
0
0
1
@suzan @andrew_trotman Can you describe what kind of recommendation papers better (or less) fit SIGIR? What better fit KDD? Honestly speaking, I don't like the saying "we kept many RecSys papers in". Sounds like RecSys papers don't naturally belong to SIGIR without your approval/mercy.
1
0
17
@suzan @HamedZamani @andrew_trotman Can you describe what kinds of Recsys papers are not “a fit to SIGIR”? I think it’s really important for Recsys researchers to consider submitting to SIGIR in future.
1
0
4
@HamedZamani @suzan @andrew_trotman Yes. Some topics of Recsys are rejected more, and I don’t quite understand the reasons of the topic discrimination, e.g., sequential recommendation…
0
0
7
@suzan @alberto_mancino @TommasoDiNoia @andrew_trotman Take the first paragraph of CFP serious, Recsys is a subset of SIGIR, so any topic of Recsys is within SIGIR’s scope. If you want to exclude some topics of Recsys, you should make it clear in CFP, before submission due. Rather than desk reject some selected papers unreasonably.
0
0
16
@suzan @andrew_trotman Right. Given the explicit statement “any aspect of information retrieval and access”, all Recsys papers are within the scope of SIGIR, regardless of which topic, e.g., sequential recommendation, rating prediction, theory, etc. No topic should be discriminated against.
1
0
7
@suzan @TommasoDiNoia @alberto_mancino @andrew_trotman "scale" is a good thing, meaning the IR problem is still attractive and valuable. "scale" should be the reason to improve the reviewing process (also the duty of PC team), rather than the execuse to reject recommendation papers...
0
0
14
@HamedZamani @suzan @andrew_trotman Not all recsys papers are desk rejected. But from my rough estimation (based on my and friends' submissions), around 30~40% recsys papers are desk rejected.
1
0
5
3
0
14
@AixinSG The second observation can be interpreted as distribution shift. P(u, i) = P(i|u) * P(u), either factor P(i|u) or P(u) changes can cause this issue.
0
0
0
@lintool Cannot agree more. But only real experts in the area can realize and appreciate the insights and power brought by your simple method.
0
0
0
RT @lintool: Reviewers automatically assume that simple is not novel. This is sheer laziness. Yes, it may be simple and obvious in retrospe…
0
494
0
RT @dlowd: Dear Reviewers: Don't reject because we made it look easy. Making it look easy is hard work. Don't reject because our solu…
0
774
0
It's a sad story...
0
0
0
"Fast MF for Online Recommendation with Implicit Feedback" accepted by #SIGIR2016 Draft&code (haven't cleaned) in:
0
5
5