Does the Milky Way have dark matter satellite galaxies? This recent paper by Weinberg et al. discusses that one potential problem with our current model of Cold Dark Matter (CDM) is that is predicts a greater number of satellite galaxies for the Milky Way than are actually observed, and satellite galaxies with larger masses than those the Milky Way has (although the paper does say that many of the Milky Way's satellite galaxies may be too dim for us to currently observe).
I know that at least dark matter galaxies are believed to have been observed (see this article).  Is there any observational or other evidence for the existence of dark matter satellite galaxies to the Milky Way?  And how have they been found?
Edit:  As has been pointed out, this is a question we likely don't have the answer to yet (otherwise there wouldn't be recent papers posing the question still), so perhaps a better question would be how could/would you find dark matter satellite galaxies of the Milky Way?
 A: This is actually a problem between simulations of structure formation and observations on a couple of different mass scales. Both galaxies and galaxy clusters appear to have nearly an order of magnitude more satellites than what is actually observed. The problem is dubbed the "missing satellite problem", or the Dwarf galaxy problem.
People have been asking the question, what if these dwarf galaxies are simply not massive enough to attract enough gas gravitationally for them to be visible? Some interesting ideas and work has been done to determine if these structures actually do exist (Yoon, Johnston, and Hogg - Clumpy Streams from Clumpy Halos: Detecting Missing Satellites with Cold Stellar Structures). Also, Beth Willman has done some interesting work on detecting the least luminous galaxies (dwarf galaxy companions to the Milky Way). In other words, it is quite possible that there are small collections of stars, dust, and gas inside these substructures that are just so faint, we can't find them unless we're looking closely.
I should also add that people have gone to lengths in the other direction, too. That is, to see how to alter the properties of dark matter in order to rid simulations of structures on that particular scale (see: Self-interacting dark matter).
The LCDM cosmological model is a very robust and well-supported model for our universe on scales of around ~1 Mpc and larger. Some discrepancies do exist, of which the missing satellite problem is but one of them.
A: Boylan-Kolchin et al. mention many of the same things that have been comments to my question:

Denser subhaloes produce a larger luminosity from dark matter annihilation.

(see fig. 5 for their simulated results),

Dark subhaloes might host at least some of the recently discovered
ultra-faint galaxies, all of which have luminosities lower
than 10$^5$ solar luminosities. Kinematic constraints favor masses and densities
for the ultra-faints that are indicative of fairly massive
subhaloes...albeit with large uncertainties at present.

So, large subhaloes may be "dark" just because they host smaller galaxies than we expect.
This paper also suggests a more gravity-oriented approach:

An alternate detection method could be through the subhaloes' tidal influence
on the MW's HI disk.

This paper by Chakrabarti et al. discusses using observed tidal disturbances in a galaxy to determine both the mass and position of the disturbing body.  From the abstract:

We describe ongoing work on a new method that allows one to approximately determine the mass and
relative position (in galactocentric radius and azimuth) of galactic companions purely from analysis of observed
disturbances in gas disks....(the galaxies we observed) span
the range from having a very low mass companion (one-hundredth the mass of the primary galaxy) to a
fairly massive companion (one-third the mass of the primary galaxy). This approach has broad implications
for many areas of astrophysics – for the indirect detection of dark matter (or dark-matter dominated dwarf
galaxies).  (emphasis added)

The mass ratio of the Milky Way to its largest satellite (the Large Magellanic Cloud) is about 100; to the next largest satellite (Small Magellanic Cloud), about 200.  So, a dark matter satellite galaxy larger than these (as predicted by the simulations of Boylan-Kolchin et al.) should be well within the detection limit of the process used in the Chakrabarti et al. paper (they were able to accurately detect a satellite with a primary galaxy/satellite ratio of 100).  With large-scale sky surveys happening as we speak, like SDSS, it may be possible to find these tidal disturbances within our own galaxy and, if a disturbance is not associated with any known satellites, we should then be able to determine the location and mass of the unseen object, which likely would be a dark matter satellite galaxy.
I think that's about as close to an answer as we can get right now, unless if there is any additional research being done that I haven't mentioned.  None of these techniques have as yet produced any definitive results (or likely haven't even been used yet for this purpose) as to whether the Milky Way has a dark matter galaxy partner.  We'll have to just wait and see.
