• Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.

RevengeTaken

Banned
Aug 12, 2018
1,711
The PlayStation 5 rumour comes from YouTuber Skullzi, who took the effort to go through some of the patents that Sony has filed. One of these patents, filed in October 2018, could indicate that the PlayStation 5 will use deep learning in order to adapt and change the experience to suit that of the user. The technology will involve the console analysing the user's playstyle, what they play, and even how they play. It will then use this information to change its UI to best reflect the user's tastes and choices.
Generally, a video game and video game assistance are adapted to a player. For example, a narrative of the video game is personalized to an experience level of the player. Similarly, assistance in interacting with a particular context of the video game is also personalised. The personalisation learns from historical interactions of players with the video game and, optionally, other video games. In an example, a deep learning neural network is implemented to generate knowledge from the historical interactions. The personalisation is set according to the knowledge.
Sonys-PlayStation-5-Might-Become-Cybernet-1.jpg


The console will even go as far as to change a game's difficulty depending on your skill level and could even decide to forgo tutorials if it calculates that your skill levels are high enough. The whole idea behind this is to give each player a fully customised and individual experience.
However, patents don't always end up being used in the final product and this could just such a case. So, it's important to take this news with a grain of salt.
 

hanmik

Editor/Writer at Popaco.dk
Verified
Oct 26, 2017
1,436
So it will be a negative, over dramatic, condescending piece of plastic?
 

ChewbieFR

Member
Nov 3, 2018
181
How cute to think this doesn't already exist in some form
You're constantly tracked because it helps selling you games you'd like and showing you more accurate ads
 

Dan Thunder

Member
Nov 2, 2017
14,009
I've always wondered about this sort of thing but more through haptics. For example having a controller that can measure heart rate, skin moisture etc and alter the difficulty or atmosphere in a game. So if you were playing a horror game but weren't feeling apprehensive the game could pick up on this through the controller and do things like throw in some creepy sound effects or make the game darker.

Something like the system described above could give Google a tremendous advantage since they bought out DeepMind a few years back. They're miles ahead of pretty much anyone in terms of AI.
 

Arapho

Member
Nov 25, 2018
71
I'm working as webmarketing manager in a national company.
Sony is probably already doing this. Every company is analything data in order to adapt the experience for the users.

Some examples were this is very visible : Netflix, Amazon
 

Tunesmith

Fraud & Player Security
Verified
Oct 25, 2017
1,936
Sony isn't the only one working on/exploring applications of adaptive AI-driven systems like those. EA for example has a team dedicated for it (SEED).
MS and other sizeable players in the industry surely has as well.
 

gofreak

Member
Oct 26, 2017
7,734
It's not so surprising... everyone is working on this stuff. I mean not just in terms of game experience, but in data mining and personalisation across your entire interaction with products, to personalise marketing for you, your UI, whatever.

Can you imagine how much data PSN generates? Of course they're going to use it to understand their users better.
 

Militaratus

The Fallen
Oct 27, 2017
1,212
I am interested in the Virtual Assistant, will this A.I. be able to take control of the second player in games like a virtual co-op partner?
 

modiz

Member
Oct 8, 2018
17,822
It's not so surprising... everyone is working on this stuff. I mean not just in terms of game experience, but in data mining and personalisation across your entire interaction with products, to personalise marketing for you, your UI, whatever.

Can you imagine how much data PSN generates? Of course they're going to use it to understand their users better.
you are right, personalized and positive user experience is seen as a very critical feature right now.
 

tokkun

Member
Oct 27, 2017
5,396
"Deep learning" is the new "blockchain", I guess.

If you want to automatically change settings based on player behavior, that's fine, but every example they give could be done programmatically. There is no need to have neural nets involved.

I'm working as webmarketing manager in a national company.
Sony is probably already doing this. Every company is analything data in order to adapt the experience for the users.

Some examples were this is very visible : Netflix, Amazon

Personalization and deep learning are not synonymous though.

Resident Evil 4 was already doing this shit back in 2005

Nope, that was done programmatically, not via deep learning.
 

DrROBschiz

Member
Oct 25, 2017
16,458
Makes me think of Last of Us 2 and that demo

Scary to think game we could be entering the next leap of Machine Learning as applied to combating player skill

"Deep learning" is the new "blockchain", I guess.

If you want to automatically change settings based on player behavior, that's fine, but every example they give could be done programmatically. There is no need to have neural nets involved.



Personalization and deep learning are not synonymous though.



Nope, that was done programmatically, not via deep learning.

I agree that its implementation shouldn't be a priority if it consumes more resources than its worth. Then again couldn't it be relegated to cloud processing?
 

gofreak

Member
Oct 26, 2017
7,734
"Deep learning" is the new "blockchain", I guess.

If you want to automatically change settings based on player behavior, that's fine, but every example they give could be done programmatically. There is no need to have neural nets involved.

It can be hard for a programmer to find the right thresholds to send code execution down one path over another.

Deep learning is just a handy way to find less obvious correlations, or use data to tweak the thresholds for when one thing kicks in over another. So yeah, you could perhaps do without it, but it might make it easier if you have the data to generate robust correlations 'automatically', and to keep updating those correlations as more data arrives, with a programmer or human overseeing things, rather than trying to figure them out themselves.