How Peter Jackson Used Artificial Intelligence to Restore the Video & Audio Featured in The Beatles: Get Back

Much has been made in recent years of the “de-aging” process­es that allow actors to cred­i­bly play char­ac­ters far younger than them­selves. But it has also become pos­si­ble to de-age film itself, as demon­strat­ed by Peter Jack­son’s cel­e­brat­ed new docu-series The Bea­t­les: Get Back. The vast major­i­ty of the mate­r­i­al that com­pris­es its near­ly eight-hour run­time was orig­i­nal­ly shot in 1969, under the direc­tion of Michael Lind­say-Hogg for the doc­u­men­tary that became Let It Be.

Those who have seen both Lin­day-Hog­g’s and Jack­son’s doc­u­men­taries will notice how much sharp­er, smoother, and more vivid the very same footage looks in the lat­ter, despite the six­teen-mil­lime­ter film hav­ing lan­guished for half a cen­tu­ry. The kind of visu­al restora­tion and enhance­ment seen in Get Back was made pos­si­ble by tech­nolo­gies that have only emerged in the past few decades — and pre­vi­ous­ly seen in Jack­son’s They Shall Not Grow Old, a doc­u­men­tary acclaimed for its restora­tion of cen­tu­ry-old World War I footage to a time-trav­el-like degree of verisimil­i­tude.

“You can’t actu­al­ly just do it with off-the-shelf soft­ware,” Jack­son explained in an inter­view about the restora­tion process­es involved in They Shall Not Grow Old. This neces­si­tat­ed mar­shal­ing, at his New Zealand com­pa­ny Park Road Post Pro­duc­tion, “a depart­ment of code writ­ers who write com­put­er code in soft­ware.” In oth­er words, a suf­fi­cient­ly ambi­tious project of visu­al revi­tal­iza­tion — mak­ing media from bygone times even more life­like than it was to begin with — becomes as much a job of tra­di­tion­al film-restora­tion or visu­al-effects as of com­put­er pro­gram­ming.

This also goes for the less obvi­ous but no-less-impres­sive treat­ment giv­en by Jack­son and his team to the audio that came with the Let It Be footage. Record­ed in large part monau­ral­ly, these tapes pre­sent­ed a for­mi­da­ble pro­duc­tion chal­lenge. John, Paul, George, and Ringo’s instru­ments share a sin­gle track with their voic­es — and not just their singing voic­es, but their speak­ing ones as well. On first lis­ten, this ren­ders many of their con­ver­sa­tions inaudi­ble, and prob­a­bly by design: “If they were in a con­ver­sa­tion,” said Jack­son, they would turn their amps up loud and they’d strum the gui­tar.”

This means of keep­ing their words from Lind­say-Hogg and his crew worked well enough in the whol­ly ana­log late 1960s, but it has proven no match for the arti­fi­cial intelligence/machine learn­ing of the 2020s. “We devised a tech­nol­o­gy that is called demix­ing,” said Jack­son. “You teach the com­put­er what a gui­tar sounds like, you teach them what a human voice sounds like, you teach it what a drum sounds like, you teach it what a bass sounds like.” Sup­plied with enough son­ic data, the sys­tem even­tu­al­ly learned to dis­tin­guish from one anoth­er not just the sounds of the Bea­t­les’ instru­ments but of their voic­es as well.

Hence, in addi­tion to Get Back’s rev­e­la­to­ry musi­cal moments, its many once-pri­vate but now crisply audi­ble exchanges between the Fab Four. “Oh, you’re record­ing our con­ver­sa­tion?” George Har­ri­son at one point asks Lind­say-Hogg in a char­ac­ter­is­tic tone of faux sur­prise. But if he could hear the record­ings today, his sur­prise would sure­ly be real.

Relat­ed Con­tent:

Watch Paul McCart­ney Com­pose The Bea­t­les Clas­sic “Get Back” Out of Thin Air (1969)

Peter Jack­son Gives Us an Entic­ing Glimpse of His Upcom­ing Bea­t­les Doc­u­men­tary The Bea­t­les: Get Back

Lennon or McCart­ney? Sci­en­tists Use Arti­fi­cial Intel­li­gence to Fig­ure Out Who Wrote Icon­ic Bea­t­les Songs

Arti­fi­cial Intel­li­gence Pro­gram Tries to Write a Bea­t­les Song: Lis­ten to “Daddy’s Car”

Watch The Bea­t­les Per­form Their Famous Rooftop Con­cert: It Hap­pened 50 Years Ago Today (Jan­u­ary 30, 1969)

How Peter Jack­son Made His State-of-the-Art World War I Doc­u­men­tary They Shall Not Grow Old: An Inside Look

Based in Seoul, Col­in Mar­shall writes and broad­casts on cities and cul­ture. His projects include the book The State­less City: a Walk through 21st-Cen­tu­ry Los Ange­les and the video series The City in Cin­e­ma. Fol­low him on Twit­ter at @colinmarshall or on Face­book.

Artificial Intelligence for Everyone: An Introductory Course from Andrew Ng, the Co-Founder of Coursera

If you fol­low edtech, you know the name Andrew Ng. He’s the Stan­ford com­put­er sci­ence pro­fes­sor who co-found­ed MOOC-provider Cours­era and lat­er became chief sci­en­tist at Baidu. Since leav­ing Baidu, he’s been work­ing on sev­er­al arti­fi­cial intel­li­gence projects, includ­ing a series of Deep Learn­ing cours­es that he unveiled in 2017. And now comes AI for Every­one–an online course that makes arti­fi­cial intel­li­gence intel­li­gi­ble to a broad audi­ence.

In this large­ly non-tech­ni­cal course, stu­dents will learn:

  • The mean­ing behind com­mon AI ter­mi­nol­o­gy, includ­ing neur­al net­works, machine learn­ing, deep learn­ing, and data sci­ence.
  • What AI real­is­ti­cal­ly can–and cannot–do.
  • How to spot oppor­tu­ni­ties to apply AI to prob­lems in your own orga­ni­za­tion.
  • What it feels like to build machine learn­ing and data sci­ence projects.
  • How to work with an AI team and build an AI strat­e­gy in an orga­ni­za­tion.
  • How to nav­i­gate eth­i­cal and soci­etal dis­cus­sions sur­round­ing AI.

The four-week course takes about eight hours to com­plete. You can audit it for free. How­ev­er if you want to earn a certificate–which you can then share on your LinkedIn pro­file, print­ed resumes and CVs–the course will run $49.

AI for Every­one will be added to our list of Free Com­put­er Sci­ence cours­es, a sub­set of our larg­er col­lec­tion, 1,700 Free Online Cours­es from Top Uni­ver­si­ties.

Relat­ed Con­tent:

Nick Cave Answers the Hot­ly Debat­ed Ques­tion: Will Arti­fi­cial Intel­li­gence Ever Be Able to Write a Great Song?

Arti­fi­cial Intel­li­gence Brings Sal­vador Dalí Back to Life: “Greet­ings, I Am Back”

Arti­fi­cial Intel­li­gence Iden­ti­fies the Six Main Arcs in Sto­ry­telling: Wel­come to the Brave New World of Lit­er­ary Crit­i­cism

New Deep Learn­ing Cours­es Released on Cours­era, with Hope of Teach­ing Mil­lions the Basics of Arti­fi­cial Intel­li­gence

by | Permalink | Make a Comment ( 4 ) |

A Free Oxford Course on Deep Learning: Cutting Edge Lessons in Artificial Intelligence

Nan­do de Fre­itas is a “machine learn­ing pro­fes­sor at Oxford Uni­ver­si­ty, a lead research sci­en­tist at Google Deep­Mind, and a Fel­low of the Cana­di­an Insti­tute For Advanced Research (CIFAR) in the Neur­al Com­pu­ta­tion and Adap­tive Per­cep­tion pro­gram.”

Above, you can watch him teach an Oxford course on Deep Learn­ing, a hot sub­field of machine learn­ing and arti­fi­cial intel­li­gence which cre­ates neur­al networks–essentially com­plex algo­rithms mod­eled loose­ly after the human brain–that can rec­og­nize pat­terns and learn to per­form tasks.

To com­ple­ment the 16 lec­tures you can also find lec­ture slides, prac­ti­cals, and prob­lems sets on this Oxford web site. If you’d like to learn about Deep Learn­ing in a MOOC for­mat, be sure to check out the new series of cours­es cre­at­ed by Andrew Ng on Cours­era.

Oxford’s Deep Learn­ing course will be added to our list of Free Online Com­put­er Sci­ence Cours­es, part of our meta col­lec­tion, 1,700 Free Online Cours­es from Top Uni­ver­si­ties.

Relat­ed Con­tent:

Google Launch­es Free Course on Deep Learn­ing: The Sci­ence of Teach­ing Com­put­ers How to Teach Them­selves

New Deep Learn­ing Cours­es Released on Cours­era, with Hope of Teach­ing Mil­lions the Basics of Arti­fi­cial Intel­li­gence

Neur­al Net­works for Machine Learn­ing: A Free Online Course

by | Permalink | Make a Comment ( 2 ) |

Google Launches a Free Course on Artificial Intelligence: Sign Up for Its New “Machine Learning Crash Course”

As part of an effort to make Arti­fi­cial Intel­li­gence more com­pre­hen­si­ble to the broad­er pub­lic, Google has cre­at­ed an edu­ca­tion­al web­site Learn with Google AI, which includes, among oth­er things, a new online course called Machine Learn­ing Crash Course. The course pro­vides “exer­cis­es, inter­ac­tive visu­al­iza­tions, and instruc­tion­al videos that any­one can use to learn and prac­tice [Machine Learn­ing] con­cepts.” To date, more than 18,000 Googlers have enrolled in the course. And now it’s avail­able for every­one, every­where. You can sup­ple­ment it with oth­er AI cours­es found in the Relat­eds below.

Machine Learn­ing Crash Course will be added to our list of Free Online Com­put­er Sci­ence Cours­es, a sub­set of our col­lec­tion, 1,700 Free Online Cours­es from Top Uni­ver­si­ties.

If you would like to sign up for Open Culture’s free email newslet­ter, please find it here. Or fol­low our posts on Threads, Face­book, BlueSky or Mastodon.

If you would like to sup­port the mis­sion of Open Cul­ture, con­sid­er mak­ing a dona­tion to our site. It’s hard to rely 100% on ads, and your con­tri­bu­tions will help us con­tin­ue pro­vid­ing the best free cul­tur­al and edu­ca­tion­al mate­ri­als to learn­ers every­where. You can con­tribute through Pay­Pal, Patre­on, and Ven­mo (@openculture). Thanks!

via Google Blog

Relat­ed Con­tent:

Arti­fi­cial Intel­li­gence: A Free Online Course from MIT

Google Launch­es Free Course on Deep Learn­ing: The Sci­ence of Teach­ing Com­put­ers How to Teach Them­selves

New Deep Learn­ing Cours­es Released on Cours­era, with Hope of Teach­ing Mil­lions the Basics of Arti­fi­cial Intel­li­gence

Neur­al Net­works for Machine Learn­ing: A Free Online Course

 

by | Permalink | Make a Comment ( 28 ) |

New Deep Learning Courses Released on Coursera, with Hope of Teaching Millions the Basics of Artificial Intelligence

FYI: If you fol­low edtech, you know the name Andrew Ng. He’s the Stan­ford com­put­er sci­ence pro­fes­sor, who co-found­ed MOOC-provider Cours­era and lat­er became chief sci­en­tist at Baidu. Since leav­ing Baidu, he’s been work­ing on three arti­fi­cial intel­li­gence projects, the first of which he unveiled yes­ter­day. On Medi­um, he wrote:

I have been work­ing on three new AI projects, and am thrilled to announce the first one: deeplearning.ai, a project ded­i­cat­ed to dis­sem­i­nat­ing AI knowl­edge, is launch­ing a new sequence of Deep Learn­ing cours­es on Cours­era. These cours­es will help you mas­ter Deep Learn­ing, apply it effec­tive­ly, and build a career in AI.

Speak­ing to the MIT Tech­nol­o­gy Review, Ng elab­o­rat­ed: “The thing that real­ly excites me today is build­ing a new AI-pow­ered soci­ety… I don’t think any one com­pa­ny could do all the work that needs to be done, so I think the only way to get there is if we teach mil­lions of peo­ple to use these AI tools so they can go and invent the things that no large com­pa­ny, or com­pa­ny I could build, could do.”

Andrew’s new 5‑part series of cours­es on Deep Learn­ing can be accessed here. Cours­es include: Neur­al Net­works and Deep Learn­ing, Improv­ing Deep Neur­al Net­works, Struc­tur­ing Machine Learn­ing Projects, Con­vo­lu­tion­al Neur­al Net­works, and Sequence Mod­els.

You can find these cours­es on our list of Free Com­put­er Sci­ence Cours­es, a sub­set of our col­lec­tion, 1,700 Free Online Cours­es from Top Uni­ver­si­ties.

If you would like to sign up for Open Culture’s free email newslet­ter, please find it here. Or fol­low our posts on Threads, Face­book, BlueSky or Mastodon.

If you would like to sup­port the mis­sion of Open Cul­ture, con­sid­er mak­ing a dona­tion to our site. It’s hard to rely 100% on ads, and your con­tri­bu­tions will help us con­tin­ue pro­vid­ing the best free cul­tur­al and edu­ca­tion­al mate­ri­als to learn­ers every­where. You can con­tribute through Pay­Pal, Patre­on, and Ven­mo (@openculture). Thanks!

Relat­ed Con­tent:

Google Launch­es Free Course on Deep Learn­ing: The Sci­ence of Teach­ing Com­put­ers How to Teach Them­selves

Google’s Deep­Mind AI Teach­es Itself to Walk, and the Results Are Kooky, No Wait, Chill­ing

Arti­fi­cial Intel­li­gence: A Free Online Course from MIT

Artificial Intelligence: A Free Online Course from MIT

Today we’re adding MIT’s course on Arti­fi­cial Intel­li­gence to our ever-grow­ing col­lec­tion, 1,700 Free Online Cours­es from Top Uni­ver­si­ties. That’s because, to para­phrase Ama­zon’s Jeff Bezos, arti­fi­cial intel­li­gence (AI) is “not just in the first inning of a long base­ball game, but at the stage where the very first bat­ter comes up.” Look around, and you will find AI everywhere–in self dri­ving cars, Siri on your phone, online cus­tomer sup­port, movie rec­om­men­da­tions on Net­flix, fraud detec­tion for your cred­it cards, etc. To be sure, there’s more to come.

Fea­tur­ing 30 lec­tures, MIT’s course “intro­duces stu­dents to the basic knowl­edge rep­re­sen­ta­tion, prob­lem solv­ing, and learn­ing meth­ods of arti­fi­cial intel­li­gence.” It includes inter­ac­tive demon­stra­tions designed to “help stu­dents gain intu­ition about how arti­fi­cial intel­li­gence meth­ods work under a vari­ety of cir­cum­stances.” And, by the end of the course, stu­dents should be able “to devel­op intel­li­gent sys­tems by assem­bling solu­tions to con­crete com­pu­ta­tion­al prob­lems; under­stand the role of knowl­edge rep­re­sen­ta­tion, prob­lem solv­ing, and learn­ing in intel­li­gent-sys­tem engi­neer­ing; and appre­ci­ate the role of prob­lem solv­ing, vision, and lan­guage in under­stand­ing human intel­li­gence from a com­pu­ta­tion­al per­spec­tive.”

Taught by Prof. Patrick Hen­ry Win­ston, the lec­tures can all be viewed above. Or watch them on YouTube and iTunes. Relat­ed course mate­ri­als (includ­ing a syl­labus) can be found on this MIT web­site. The text­book, avail­able on Ama­zon, was writ­ten by Pro­fes­sor Win­ston.

If you would like to sign up for Open Culture’s free email newslet­ter, please find it here. Or fol­low our posts on Threads, Face­book, BlueSky or Mastodon.

If you would like to sup­port the mis­sion of Open Cul­ture, con­sid­er mak­ing a dona­tion to our site. It’s hard to rely 100% on ads, and your con­tri­bu­tions will help us con­tin­ue pro­vid­ing the best free cul­tur­al and edu­ca­tion­al mate­ri­als to learn­ers every­where. You can con­tribute through Pay­Pal, Patre­on, and Ven­mo (@openculture). Thanks!

Relat­ed Con­tent:

Free Online Com­put­er Sci­ence Cours­es

Hayao Miyaza­ki Tells Video Game Mak­ers What He Thinks of Their Char­ac­ters Made with Arti­fi­cial Intel­li­gence: “I’m Utter­ly Dis­gust­ed. This Is an Insult to Life Itself”

Arti­fi­cial Intel­li­gence Pro­gram Tries to Write a Bea­t­les Song: Lis­ten to “Daddy’s Car”

Two Arti­fi­cial Intel­li­gence Chat­bots Talk to Each Oth­er & Get Into a Deep Philo­soph­i­cal Con­ver­sa­tion

Noam Chom­sky Explains Where Arti­fi­cial Intel­li­gence Went Wrong

by | Permalink | Make a Comment ( 26 ) |

« Go Back
Quantcast