This computer is learning to read your mind | DIY Neuroscience, a TED series

Translator: Joseph Geni Reviewer: Krystian Aparta

Translator: Nauka N. Prasadini Reviewer: annisa putri

Membaca pikiran. Anda melihatnya di film sci-fi:

mesin yang dapat membaca pikiran kita.

Namun, saat ini, ada alat

yang dapat membaca aktivitas listrik otak kita.

Kita menyebut alat itu EEG.

Apakah ada informasi dalam gelombang otak ini?

Greg Gage: Mind-reading. You've seen this in sci-fi movies:

Jika ada, apakah kita bisa melatih komputer membaca pikiran?

machines that can read our thoughts.

However, there are devices today

Rekan saya Nathan sedang mencoba meretas mesin EEG

that can read the electrical activity from our brains.

untuk membuat mesin pembaca pikiran.

We call this the EEG.

[DIY Neurosains]

Is there information contained in these brainwaves?

Beginilah cara kerja EEG.

And if so, could we train a computer to read our thoughts?

Terdapat otak di kepala Anda,

dan otak tersebut terdiri dari milyaran neuron.

My buddy Nathan has been working to hack the EEG

Masing-masing neuron saling mengirimkan pesan elektrik.

to build a mind-reading machine.

Pesan kecil ini dapat digabungkan menjadi gelombang elektrik

[DIY Neuroscience]

So this is how the EEG works.

yang kita deteksi di layar.

Inside your head is a brain,

Pada prinsipnya, EEG dapat menunjukkan hal-hal berskala besar,

and that brain is made out of billions of neurons.

misalnya saat Anda sedang tidur atau terjaga.

Each of those neurons sends an electrical message to each other.

Bisakah ia menunjukkan hal lain?

Apakah ia sungguh membaca pikiran?

These small messages can combine to make an electrical wave

Kita akan mengujinya,

that we can detect on a monitor.

dan kita tidak akan mulai dari ide-ide rumit.

Now traditionally, the EEG can tell us large-scale things,

Namun dari yang sederhana.

Dapatkah kita menafsir penglihatan orang

for example if you're asleep or if you're alert.

hanya dari gelombang otaknya?

But can it tell us anything else?

Nathan memulai dengan menempelkan elektroda ke kepala Christy.

Can it actually read our thoughts?

We're going to test this,

Nathan: Hidupku rasanya kusut.

and we're not going to start with some complex thoughts.

(tertawa)

Greg Gage: Dia akan memperlihatkan Christy gambar

We're going to do something very simple.

Can we interpret what someone is seeing using only their brainwaves?

dari empat kategori berbeda.

Nathan: wajah, rumah, pemandangan, dan gambar aneh.

Nathan's going to begin by placing electrodes on Christy's head.

GG: Saat kami menunjukkan Christy ratusan gambar tadi,

Nathan: My life is tangled.

kami juga menangkap gelombang elektrik ke komputer Nathan.

(Laughter)

GG: And then he's going to show her a bunch of pictures

Kami ingin mendeteksi

from four different categories.

adanya informasi visual foto tersebut

Nathan: Face, house, scenery and weird pictures.

yang dikandung gelombang otak,

jadi saat selesai, kami akan melihat apakah EEG

GG: As we show Christy hundreds of these images,

dapat memberi tahu jenis gambar apa yang Christy lihat,

we are also capturing the electrical waves onto Nathan's computer.

dan jika terjadi, tiap kategori seharusnya memicu sinyal otak yang berbeda.

We want to see if we can detect any visual information about the photos

OK, setelah kami mengumpulkan semua data mentah EEG,

contained in the brainwaves,

so when we're done, we're going to see if the EEG

inilah hasilnya.

Ini terlihat berantakan. Mari kita susun berdasarkan gambar.

can tell us what kind of picture Christy is looking at,

and if it does, each category should trigger a different brain signal.

Nah, ini masih terlihat riuh untuk mencari perbedaannya,

tetapi jika kita merata-rata EEG di semua jenis gambar

OK, so we collected all the raw EEG data,

dengan menyelaraskannya saat gambar muncul,

and this is what we got.

It all looks pretty messy, so let's arrange them by picture.

kita bisa menghilangkan derau ini

dan nantinya, kita bisa melihat pola dominan

Now, still a bit too noisy to see any differences,

muncul dari tiap kategori.

Sekarang, semua sinyal masih sangat mirip.

but if we average the EEG across all image types

Mari kita cermati.

by aligning them to when the image first appeared,

Sekitar seratus milidetik setelah gambar muncul,

kami melihat lonjakan positif dalam keempat kasus,

we can remove this noise,

and pretty soon, we can see some dominant patterns

dan kami menyebutnya P100, dan kami berpendapat

emerge for each category.

Now the signals all still look pretty similar.

inilah yang terjadi di otak Anda saat mengenali sebuah benda.

Let's take a closer look.

Tapi astaga, coba lihat sinyal wajah.

About a hundred milliseconds after the image comes on,

Terlihat berbeda dari yang lain.

we see a positive bump in all four cases,

Ada penurunan negatif sekitar 170 milidetik

and we call this the P100, and what we think that is

setelah gambarnya muncul.

Apa penyebab terjadinya hal ini?

is what happens in your brain when you recognize an object.

Penelitian menunjukkan otak kita punya banyak neuron yang bertugas

But damn, look at that signal for the face.

mengenali wajah manusia,

It looks different than the others.

jadi lonjakan N170 ini bisa jadi karena neuron-neuron itu

There's a negative dip about 170 milliseconds

bekerja sekaligus di posisi yang sama,

after the image comes on.

dan kita bisa mendeteksinya di EEG.

What could be going on here?

Research shows that our brain has a lot of neurons that are dedicated

Sekarang ada dua kesimpulan.

Satu, mata kita tak bisa benar-benar mendeteksi perbedaan pola

to recognizing human faces,

so this N170 spike could be all those neurons

tanpa merata-ratakan deraunya,

dan dua, bahkan setelah menghilangkan deraunya,

firing at once in the same location,

mata kita hanya bisa menangkap sinyal yang terkait wajah.

and we can detect that in the EEG.

Di sinilah kita beralih ke machine learning.

So there are two takeaways here.

One, our eyes can't really detect the differences in patterns

Mata kita tidak terampil dalam menemukan pola pada derau data,

without averaging out the noise,

and two, even after removing the noise,

namun algoritma machine learning dirancang melakukan hal itu,

our eyes can only pick up the signals associated with faces.

jadi bisakah kita mengambil berbagai foto dan gambar

So this is where we turn to machine learning.

lalu mengumpankannya dan melatih komputer

Now, our eyes are not very good at picking up patterns in noisy data,

agar dapat menafsirkan apa yang Christy lihat di saat yang sama?

but machine learning algorithms are designed to do just that,

Kami mencoba untuk mengkode informasi yang keluar dari EEG-nya

so could we take a lot of pictures and a lot of data

di saat yang sama

dan memperkirakan apa yang dilihatnya.

and feed it in and train a computer

Jika berhasil, yang perlu dilihat

to be able to interpret what Christy is looking at in real time?

adalah tiap kali dia mendapat gambar pemandangan,

We're trying to code the information that's coming out of her EEG

mesinnya harus menyebut pemandangan, pemandangan.

Wajah -- wajah, wajah, wajah, wajah,

in real time

tetapi yang kita temukan justru tidak bekerja seperti itu.

and predict what it is that her eyes are looking at.

And if it works, what we should see

is every time that she gets a picture of scenery,

it should say scenery, scenery, scenery, scenery.

(tertawa)

A face -- face, face, face, face,

OK.

but it's not quite working that way, is what we're discovering.

Sutradara: Ada apa? GG: Kurasa kita perlu pekerjaan baru.

(tertawa)

OK, ternyata yang tadi gagal besar.

(Laughter)

Tapi kami masih penasaran:

Seberapa jauh kita bisa dorong teknologi ini?

OK.

Kami meninjau ulang usaha kami.

Kami sadar bahwa data masuk dengan sangat cepat ke komputer kami,

Director: So what's going on here? GG: We need a new career, I think.

(Laughter)

tanpa adanya rentang kapan gambar itu muncul,

OK, so that was a massive failure.

dan itu seperti membaca kalimat yang sangat panjang

But we're still curious: How far could we push this technology?

tanpa spasi pada kata-katanya.

Ini akan sulit dibaca,

And we looked back at what we did.

namun jika kita tambahkan spasi, kata individual akan muncul

We noticed that the data was coming into our computer very quickly,

without any timing of when the images came on,

dan kalimat menjadi mudah dipahami.

and that's the equivalent of reading a very long sentence

Namun bagaimana jika kita mengakalinya?

Dengan sensor, kita bisa melihat saat gambar muncul di komputer.

without spaces between the words.

It would be hard to read,

Dengan begitu, gelombang otak berhenti menjadi aliran informasi tanpa henti,

but once we add the spaces, individual words appear

dan sebagai gantinya, menjadi kumpulan makna individual.

and it becomes a lot more understandable.

But what if we cheat a little bit?

Kita juga akan sedikit mengakalinya,

By using a sensor, we can tell the computer when the image first appears.

dengan membatasi kategori jadi dua.

Lihat jika kita bisa membaca pikiran langsung.

That way, the brainwave stops being a continuous stream of information,

Dalam percobaan baru ini,

kami akan sedikit menyempitkannya

and instead becomes individual packets of meaning.

sehingga kita tahu permulaan gambarnya

Also, we're going to cheat a little bit more,

dan kita akan membatasi kategori ke “wajah” atau “pemandangan.”

by limiting the categories to two.

Let's see if we can do some real-time mind-reading.

Nathan: Wajah. Benar.

In this new experiment,

Pemandangan. Benar.

we're going to constrict it a little bit more

GG: Nah sekarang, tiap kali ada gambar muncul,

so that we know the onset of the image

kita akan memotret awal gambar

and we're going to limit the categories to "face" or "scenery."

dan menguraikan EEG-nya.

Mesinnya mulai benar.

Nathan: Face. Correct.

Nathan: Ya. Wajah. Benar.

Scenery. Correct.

GGL: Jadi ada informasi di sinyal EEG. Keren sekali.

GG: So right now, every time the image comes on,

Kita hanya harus menyelaraskannya dengan awal gambar.

we're taking a picture of the onset of the image

Nathan: Pemandangan. Benar.

and decoding the EEG.

It's getting correct.

Nathan: Yes. Face. Correct.

Wajah. Oke.

GG: Artinya ada informasi di balik itu,

GG: So there is information in the EEG signal, which is cool.

yang berarti jika kita tahu kapan gambar akan muncul,

We just had to align it to the onset of the image.

kita bisa tahu apa kategori gambar itu,

Nathan: Scenery. Correct.

mungkin, setidaknya secara umum, dengan melihat potensi yang timbul.

Face. Yeah.

GG: This means there is some information there,

Nathan: Tepat sekali.

so if we know at what time the picture came on,

GG: Andai kau bilang di awal proyek, ini mungkin terjadi,

we can tell what type of picture it was,

Aku akan bilang mustahil.

Aku tak mengira kita akan berhasil.

possibly, at least on average, by looking at these evoked potentials.

Apakah kami berhasil membaca pikiran?

Ya, tapi kami terpaksa curang.

Nathan: Exactly.

Ternyata Anda bisa menemukan hal-hal menarik dari EEG,

GG: If you had told me at the beginning of this project this was possible,

misalnya saat Anda melihat wajah orang,

I would have said no way.

namun masih ada berbagai keterbatasan.

I literally did not think we could do this.

Mungkin kemajuan machine learning akan membuat langkah besar,

Did our mind-reading experiment really work?

Yes, but we had to do a lot of cheating.

dan nantinya, kita dapat menguraikan isi pikiran kita.

It turns out you can find some interesting things in the EEG,

Tapi untuk saat ini,

for example if you're looking at someone's face,

jika ada perusahaan yang mengklaim memanfaatkan gelombang otak

but it does have a lot of limitations.

untuk mengendalikan perangkat,

Perhaps advances in machine learning will make huge strides,

Anda berhak meragukannya.

and one day we will be able to decode what's going on in our thoughts.

But for now, the next time a company says that they can harness your brainwaves

to be able to control devices,

it is your right, it is your duty to be skeptical.