Artwork

ืชื•ื›ืŸ ืžืกื•ืคืง ืขืœ ื™ื“ื™ TWIML and Sam Charrington. ื›ืœ ืชื•ื›ืŸ ื”ืคื•ื“ืงืืกื˜ื™ื ื›ื•ืœืœ ืคืจืงื™ื, ื’ืจืคื™ืงื” ื•ืชื™ืื•ืจื™ ืคื•ื“ืงืืกื˜ื™ื ืžื•ืขืœื™ื ื•ืžืกื•ืคืงื™ื ื™ืฉื™ืจื•ืช ืขืœ ื™ื“ื™ TWIML and Sam Charrington ืื• ืฉื•ืชืฃ ืคืœื˜ืคื•ืจืžืช ื”ืคื•ื“ืงืืกื˜ ืฉืœื•. ืื ืืชื” ืžืืžื™ืŸ ืฉืžื™ืฉื”ื• ืžืฉืชืžืฉ ื‘ื™ืฆื™ืจื” ืฉืœืš ื”ืžื•ื’ื ืช ื‘ื–ื›ื•ื™ื•ืช ื™ื•ืฆืจื™ื ืœืœื ืจืฉื•ืชืš, ืืชื” ื™ื›ื•ืœ ืœืขืงื•ื‘ ืื—ืจ ื”ืชื”ืœื™ืš ื”ืžืชื•ืืจ ื›ืืŸ https://he.player.fm/legal.
Player FM - ืืคืœื™ืงืฆื™ื™ืช ืคื•ื“ืงืืกื˜
ื”ืชื—ืœ ื‘ืžืฆื‘ ืœื ืžืงื•ื•ืŸ ืขื ื”ืืคืœื™ืงืฆื™ื” Player FM !

Can Language Models Be Too Big? ๐Ÿฆœ with Emily Bender and Margaret Mitchell - #467

54:02
 
ืฉืชืคื•
 

Manage episode 288308622 series 2355587
ืชื•ื›ืŸ ืžืกื•ืคืง ืขืœ ื™ื“ื™ TWIML and Sam Charrington. ื›ืœ ืชื•ื›ืŸ ื”ืคื•ื“ืงืืกื˜ื™ื ื›ื•ืœืœ ืคืจืงื™ื, ื’ืจืคื™ืงื” ื•ืชื™ืื•ืจื™ ืคื•ื“ืงืืกื˜ื™ื ืžื•ืขืœื™ื ื•ืžืกื•ืคืงื™ื ื™ืฉื™ืจื•ืช ืขืœ ื™ื“ื™ TWIML and Sam Charrington ืื• ืฉื•ืชืฃ ืคืœื˜ืคื•ืจืžืช ื”ืคื•ื“ืงืืกื˜ ืฉืœื•. ืื ืืชื” ืžืืžื™ืŸ ืฉืžื™ืฉื”ื• ืžืฉืชืžืฉ ื‘ื™ืฆื™ืจื” ืฉืœืš ื”ืžื•ื’ื ืช ื‘ื–ื›ื•ื™ื•ืช ื™ื•ืฆืจื™ื ืœืœื ืจืฉื•ืชืš, ืืชื” ื™ื›ื•ืœ ืœืขืงื•ื‘ ืื—ืจ ื”ืชื”ืœื™ืš ื”ืžืชื•ืืจ ื›ืืŸ https://he.player.fm/legal.

Today weโ€™re joined by Emily M. Bender, Professor at the University of Washington, and AI Researcher, Margaret Mitchell.

Emily and Meg, as well as Timnit Gebru and Angelina McMillan-Major, are co-authors on the paper On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? ๐Ÿฆœ. As most of you undoubtedly know by now, there has been much controversy surrounding, and fallout from, this paper. In this conversation, our main priority was to focus on the message of the paper itself. We spend some time discussing the historical context for the paper, then turn to the goals of the paper, discussing the many reasons why the ever-growing datasets and models are not necessarily the direction we should be going.

We explore the cost of these training datasets, both literal and environmental, as well as the bias implications of these models, and of course the perpetual debate about responsibility when building and deploying ML systems. Finally, we discuss the thin line between AI hype and useful AI systems, and the importance of doing pre-mortems to truly flesh out any issues you could potentially come across prior to building models, and much much more.

The complete show notes for this episode can be found at twimlai.com/go/467.

  continue reading

695 ืคืจืงื™ื

Artwork
iconืฉืชืคื•
 
Manage episode 288308622 series 2355587
ืชื•ื›ืŸ ืžืกื•ืคืง ืขืœ ื™ื“ื™ TWIML and Sam Charrington. ื›ืœ ืชื•ื›ืŸ ื”ืคื•ื“ืงืืกื˜ื™ื ื›ื•ืœืœ ืคืจืงื™ื, ื’ืจืคื™ืงื” ื•ืชื™ืื•ืจื™ ืคื•ื“ืงืืกื˜ื™ื ืžื•ืขืœื™ื ื•ืžืกื•ืคืงื™ื ื™ืฉื™ืจื•ืช ืขืœ ื™ื“ื™ TWIML and Sam Charrington ืื• ืฉื•ืชืฃ ืคืœื˜ืคื•ืจืžืช ื”ืคื•ื“ืงืืกื˜ ืฉืœื•. ืื ืืชื” ืžืืžื™ืŸ ืฉืžื™ืฉื”ื• ืžืฉืชืžืฉ ื‘ื™ืฆื™ืจื” ืฉืœืš ื”ืžื•ื’ื ืช ื‘ื–ื›ื•ื™ื•ืช ื™ื•ืฆืจื™ื ืœืœื ืจืฉื•ืชืš, ืืชื” ื™ื›ื•ืœ ืœืขืงื•ื‘ ืื—ืจ ื”ืชื”ืœื™ืš ื”ืžืชื•ืืจ ื›ืืŸ https://he.player.fm/legal.

Today weโ€™re joined by Emily M. Bender, Professor at the University of Washington, and AI Researcher, Margaret Mitchell.

Emily and Meg, as well as Timnit Gebru and Angelina McMillan-Major, are co-authors on the paper On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? ๐Ÿฆœ. As most of you undoubtedly know by now, there has been much controversy surrounding, and fallout from, this paper. In this conversation, our main priority was to focus on the message of the paper itself. We spend some time discussing the historical context for the paper, then turn to the goals of the paper, discussing the many reasons why the ever-growing datasets and models are not necessarily the direction we should be going.

We explore the cost of these training datasets, both literal and environmental, as well as the bias implications of these models, and of course the perpetual debate about responsibility when building and deploying ML systems. Finally, we discuss the thin line between AI hype and useful AI systems, and the importance of doing pre-mortems to truly flesh out any issues you could potentially come across prior to building models, and much much more.

The complete show notes for this episode can be found at twimlai.com/go/467.

  continue reading

695 ืคืจืงื™ื

All episodes

×
 
Loading …

ื‘ืจื•ื›ื™ื ื”ื‘ืื™ื ืืœ Player FM!

Player FM ืกื•ืจืง ืืช ื”ืื™ื ื˜ืจื ื˜ ืขื‘ื•ืจ ืคื•ื“ืงืืกื˜ื™ื ื‘ืื™ื›ื•ืช ื’ื‘ื•ื”ื” ื‘ืฉื‘ื™ืœื›ื ื›ื“ื™ ืฉืชื”ื ื• ืžื”ื ื›ืจื’ืข. ื–ื” ื™ื™ืฉื•ื ื”ืคื•ื“ืงืืกื˜ ื”ื˜ื•ื‘ ื‘ื™ื•ืชืจ ื•ื”ื•ื ืขื•ื‘ื“ ืขืœ ืื ื“ืจื•ืื™ื“, iPhone ื•ืื™ื ื˜ืจื ื˜. ื”ื™ืจืฉืžื• ืœืกื ื›ืจื•ืŸ ืžื ื•ื™ื™ื ื‘ืžื›ืฉื™ืจื™ื ืฉื•ื ื™ื.

 

ืžื“ืจื™ืš ืขื–ืจ ืžื”ื™ืจ

ืคื•ื“ืงืืกื˜ื™ื ืžื•ื‘ื™ืœื™ื