8 ms·
defendants should always be able to challenge expert witnesses qualifications. and how could anyone on hacker news claim there is 'no evidence' that unreviewed
by josephorjoe 6y ago
defendants should always be able to challenge expert witnesses qualifications.
and how could anyone on hacker news claim there is 'no evidence' that unreviewed software may do something other than what people claim it does? especially if it 170,000 lines of matlab?
from the article:
> "Without scrutinizing its software's source code – a human-made set of instructions that may contain bugs, glitches, and defects – in the context of an adversarial system, no finding that it properly implements the underlying science could realistically be made," the ruling says.
- x0x0 6y agoParticularly when TrueAllele specializes in >probabilistic genotyping [1] and >solves mixed DNA samples, low template DNA and kinship problems [1] And the product sheet discusses > models for PCR artifacts [2] This sure feels like they go way beyond typical dna match technology into maximally extracting information via statistics, so their methods definitely should be open for review. [1] https://www.cybgen.com/products/casework.shtml https://www.cybgen.com/products/casework.shtml [2] https://www.cybgen.com/solutions/brochures/unique_features.pdf https://www.cybgen.com/solutions/brochures/unique_features.p...