Deep see exploration is already carried out by robots, and how do you put humans in a suns corona or deep inside the atmosphere of jupiter. Throwing out numbers like humans are 100 to 150 times more productive than robots is absurd and false. We already have to use robotic tools to extend our senses in many environments (i.e. like deep see that i've already referenced). The idea that we can somehow get more information by having our eyes somewhere rather than tools that can see wavelengths that extend orders of magnitudes beyond our eyes is just not true. All the productivity gain by human is based on the false assumption that humans will be these vast amplifiers of rapid decision making. Well I think that assumption is greatly flawed. Humans plans exploration agendas and there is a hierarchy of decision making. Some aspects of the hierarchy of an exploration campaign might be:

the Mars site to be explored, the goals of the exploration
areas within the site, to have more detailed exploration
month, week, and day plans
specific operations to be performed hour by hour
The actual application of insturments and tools to perform specific operations

Looking at the above list, none of them are impacted productivity wise by human presence other than the last.
Robotic improvements to be able to give fairly general commands (e.g. examine the rock, drill a 5 ft core sample), are in the offing so that the last issue wouldn't be greatly aided by human presence either. Increasing information bandwidth back to earth by a few orders of magnitude (easily doable), coupled with next generation improvements in robotics (expert system planners, increased mobility) will wipe out most of the presumed productivity advantage humans have over robots.

See my other post for other items and reasons robotic exploration is much more effective than human, and the for when we should do manned exploration.