Specific conference websites to find out titles and typically the author's personal site.
That's mostly what i've been doing too. But it's hard. for the new conferences, the article usually isn't available online. and for the older instances, the conferences' websites sometimes lack the history.
eprintweb: http://eprintweb.org/S/rss/cs in google reader. There are a few of the math/ and q-bio/q-fin feeds that are usually interesting reading.
I'm not sure why I didn't use the arXiv RSS feeds directly.