黑料吃瓜资源

Artificial intelligence has entered our daily lives. First, it was ChatGPT. Now, it鈥檚 AI-generated pizza and beer commercials. While we can鈥檛 trust AI to be perfect, it turns out that sometimes we can鈥檛 trust ourselves with AI either.

黑料吃瓜资源 (CSHL) Assistant Professor Peter Koo has found that scientists using popular computational tools to interpret AI predictions are picking up too much 鈥渘oise,鈥� or extra information, when analyzing DNA. And he鈥檚 found a way to fix this. Now, with just a couple new lines of code, scientists can get more reliable explanations out of powerful AIs known as deep neural networks. That means they can continue chasing down genuine DNA features. Those features might just signal the next breakthrough in health and medicine. But scientists won鈥檛 see the signals if they鈥檙e drowned out by too much noise.

So, what causes the meddlesome noise? It鈥檚 a mysterious and invisible source like digital 鈥渄ark matter.鈥� Physicists and astronomers believe most of the universe is filled with dark matter, a material that exerts gravitational effects but that no one has yet seen. Similarly, Koo and his team discovered the data that AI is being trained on lacks critical information, leading to significant blind spots. Even worse, those blind spots get factored in when interpreting AI predictions of DNA function. Koo says:

鈥淭he deep neural network is incorporating this random behavior because it learns a function everywhere. But DNA is only in a small subspace of that. And it introduces a lot of noise. And so we show that this problem actually does introduce a lot of noise across a wide variety of prominent AI models.鈥�

The digital dark matter is a result of scientists borrowing computational techniques from computer vision AI. DNA data, unlike images, is confined to a combination of four nucleotide letters: A, C, G, T. But image data in the form of pixels can be long and continuous. In other words, we鈥檙e feeding AI an input it doesn鈥檛 know how to handle properly.

By applying Koo鈥檚 computational correction, scientists can interpret AI鈥檚 DNA analyses more accurately. He says:

鈥淲e end up seeing sites that become much more crisp and clean, and there is less spurious noise in other regions. One-off nucleotides that are deemed to be very important all of a sudden disappear.鈥�

Koo believes noise disturbance affects more than AI-powered DNA analyzers. He thinks it鈥檚 a widespread affliction among computational processes involving similar types of data. Remember, dark matter is everywhere. Thankfully, Koo鈥檚 new tool can help bring scientists out of the darkness and into the light.

Written by: Luis Sandoval, Communications Specialist | [email protected] | 516-367-6826

Funding

National Institutes of Health, Simons Center for Quantitative Biology at 黑料吃瓜资源

Citation

Majdandzic, A., et al., 鈥淐orrecting gradient-based interpretations of deep neural networks for genomics鈥�, Genome Biology, May 9, 2023. DOI: