Re: What criteria unambiguously define biological evolution?

Unfortunately, biology is fraught with seeming exceptions and ambiguities, because the most accurate description of any biological process is that it follows the rules of chemistry and physics. This holds for genetics, biochemistry, and evolution. For instance, you mentioned the "Central Dogma of Microbiology" (which is now the "Rule-of-Thumb of High School Biology"), which is such an oversimplification of the interactions between nucleic acids and proteins that most geneticists (including Watson and Crick) abandoned it back in the late 70's. Evolution is far trickier to pin down, again because the interactions between genes, individuals, populations, and species can't really be summarized in one sentence. That said, here's my best shot:

Biological evolution is the change in organisms over time, often in response to 
natural selection,  such that  the descendants are genotypically different from 
their ancestors.

Notice that there is no mention of complexity or benefit; while these factors often accompany evolution, they are not necessary for evolution to occur. All that matters for a process to be considered evolutionary is genetic change. That said, I feel compelled to address the issues of complexity and benefit, to show their disconnection with evolution.

First, complexity: To say that increasing complexity is a hallmark of evolution suggests that for any evolutionary process, the descendent organisms are more complex that their ancestors. However, if we look at the beginning of human evolution, specifically at the separation of the hominid lineage from the panid (chimps and bonobos) lineage, we see a diminution in chromosome number, resulting from the fusion of two ape chromosomes to form a single hominid chromosome - rather than being more complex, the human karyotype is simpler than our ape ancestors'. In fact, there are many instances of changes in chromosome number between closely related species (both increases and decreases), in which these changes were neutral in terms of fitness, but created a reproductive boundary as part of speciation. In plants, there are several instances of tetraploidy (having 4 sets of chromosomes instead of just 2), in which the descendent plants appear hardier than their ancestors - though this probably has more to do with hybrid strength than some value in complexity. Although more complex than their predecessors, many of these plants are still outcompeted for resources by diploid (less complex) weeds. Humans are much more complex than bacteria and archaea, yet they are able to occupy niches that have required recent technological achievements for humans to even detect. In fact, bacterial evolution drives more toward streamlining and high efficiency than complexity, and bacteria are by far the best adapted to living in every possible corner of this planet.

Next, benefit: Evolution does not require natural selection to occur, i.e. a genetic change in an organism does not have to be beneficial for it to be considered evolution. Speciation can occur simply through genetic drift in isolated populations, even when the selective pressures on those populations are identical. As mentioned above, simply altering the arrangement of the chromosomes of an organism without altering its genes can form a reproductive boundary between populations such that they must be considered separate species, even though there is no difference in fitness between the populations. That is not to say that natural selection is meaningless. On the contrary, natural selection has been the major motivational force behind the evolution of life on Earth. But, the fact that genetic changes can occur in an organism that are neutral regarding the fitness of the organism and still lead to speciation, suggest that a genetic change need not be beneficial to be evolutionarily important.

I hope this answer wasn't too disheartening, but to iterate what I said at the beginning, the most accurate description of any biological process is that it follows the rules of chemistry and physics. As long as it doesn't violate those rules, anything can happen.