Toggle light / dark theme

Unifying Language Learning Paradigms

Existing pre-trained models are generally geared towards a particular class of problems. To date, there seems to be still no consensus on what the right architecture and pre-training setup should be. This paper presents a unified framework for pre-training models that are universally effective across datasets and setups. We begin by disentangling architectural archetypes with pre-training objectives — two concepts that are commonly conflated. Next, we present a generalized and unified perspective for self-supervision in NLP and show how different pre-training objectives can be cast as one another and how interpolating between different objectives can be effective. We then propose Mixture-of-Denoisers (MoD), a pre-training objective that combines diverse pre-training paradigms together.

Deep Under the Antarctic Ice, Scientists Discover Vast Reservoir of Ancient Water

A vast reservoir of ancient water has been found thousands of feet under the ice in western Antarctica, scientists said in a paper published Thursday in the journal Science.

Researchers had long suspected but never before established the existence of such hidden pockets of Antarctic groundwater, which they believe act to lessen friction between ice sheets and underlying bedrock to make the ice more prone to slide from the continent’s interior toward the surrounding ocean.

Meet some of the oldest “undead” spacecraft that are still going strong

Time will tell if more effective strategies can be developed to manage space junk in the future. But, as you are about to find out, we may not want to clear up space entirely.

Some of these “dead” spacecraft may still function!

1. Voyager 1 and 2 are still going strong.

Perhaps the most famous example of old spacecraft still in use today are Voyager 1 and 2. By far the farthest-traveled human-made objects ever sent into space, these amazing pieces of kit are still faithfully sending data back to Earth.

New study investigates photonics for artificial intelligence and neuromorphic computing

Scientists have given a fascinating new insight into the next steps to develop fast, energy-efficient, future computing systems that use light instead of electrons to process and store information—incorporating hardware inspired directly by the functioning of the human brain.

A team of scientists, including Professor C. David Wright from the University of Exeter, has explored the future potential for computer systems by using photonics in place of conventional electronics.

The article is published today (January 29th 2021) in the prestigious journal Nature Photonics.

/* */