mirror of
https://github.com/Brandon-Rozek/website.git
synced 2024-11-21 15:56:29 -05:00
Typo fixing
This commit is contained in:
parent
cc8b4d90e1
commit
d06aa84707
11 changed files with 20 additions and 20 deletions
|
@ -20,7 +20,7 @@ u_\mathfrak{a}(s) \rightarrow \mathbb{R}
|
|||
$$
|
||||
|
||||
If the utility is positive, then we say that there are more benefits than downsides. If negative, then vice versa. As differing agents have different preferences,
|
||||
the utility function is dependent upon the agent $\mathfrak{a}$. The input of the function is some *state* $s$ which represents the environment that the agent is in. The output is a continous real number.
|
||||
the utility function is dependent upon the agent $\mathfrak{a}$. The input of the function is some *state* $s$ which represents the environment that the agent is in. The output is a continuous real number.
|
||||
|
||||
A rational action in this context is the one that provides the highest utility in the next state.
|
||||
$$
|
||||
|
@ -51,7 +51,7 @@ Russ Roberts didn't mention MDPs in his book, but I think he had something simil
|
|||
|
||||
Russ then remarks on one of his key points.
|
||||
|
||||
> One of the weidest parts of the decision, as Paul pints out, is that once you become a vampire, what you like and what you dislike change.
|
||||
> One of the weirdest parts of the decision, as Paul pints out, is that once you become a vampire, what you like and what you dislike change.
|
||||
|
||||
Let's look at it in terms of value iteration. The value of becoming a vampire is equal to the reward during the initial transformation plus the discounted value of living life after the transformation.
|
||||
$$
|
||||
|
@ -71,7 +71,7 @@ Given the formalism, what are some possible reasons why we don't know this?
|
|||
|
||||
## Not Knowing the Action
|
||||
|
||||
When an agent performs the action $a_{transform}$ does it know that it performed that specific action? One way of looking at this is to see if the agent is able to tell apart any two arbitrary actions, say $a_{tranform}$ and $a_{morph}$. I don't believe this problem is well studied in literature. Do let me know if you know of any literature covering an agent confusing actions.
|
||||
When an agent performs the action $a_{transform}$ does it know that it performed that specific action? One way of looking at this is to see if the agent is able to tell apart any two arbitrary actions, say $a_{transform}$ and $a_{morph}$. I don't believe this problem is well studied in literature. Do let me know if you know of any literature covering an agent confusing actions.
|
||||
|
||||
In the case of the vampire problem, it's likely safe to assume that the human knows the transformation action.
|
||||
|
||||
|
@ -106,7 +106,7 @@ of the other party may not line up exactly with yours.
|
|||
|
||||
## Partial Observability
|
||||
|
||||
Are we able to perceive the entirity of
|
||||
Are we able to perceive the entirety of
|
||||
a state? If not, within the field of
|
||||
sequential decision making, we call the
|
||||
problem *partially observable*.
|
||||
|
@ -117,7 +117,7 @@ limiting visibility of other traffic.
|
|||
- Group Decision Making: Agents often don't have insight to other's thought processes or perceptions.
|
||||
|
||||
|
||||
One popular way of modelling this is through
|
||||
One popular way of modeling this is through
|
||||
the introduction of belief states. A belief
|
||||
state encapsulates the set of possible states.
|
||||
For example, let's say that I prefer heads on a
|
||||
|
@ -133,7 +133,7 @@ The hope is that by keeping track of the
|
|||
list of possible states, when we are interacting
|
||||
with the environment, we can filter out impossible
|
||||
transitions until we get to a single possibility.
|
||||
However, there's no guarentee that this will occur.
|
||||
However, there's no guarantee that this will occur.
|
||||
Also the initial belief state may be infinitely long.
|
||||
|
||||
Another way that we can hope to tackle this issue
|
||||
|
|
|
@ -5,7 +5,7 @@ draft: false
|
|||
tags: ["Virtualization"]
|
||||
---
|
||||
|
||||
Traditionally I would hold one linux distribution per USB drive. Though with 32 GB USB Flash drives only costing $12, why hold only one per drive? That's where Ventoy comes in. With Ventoy, we can store multiple ISOs which it then presents as a boot screen.
|
||||
Traditionally I would hold one Linux distribution per USB drive. Though with 32 GB USB Flash drives only costing $12, why hold only one per drive? That's where Ventoy comes in. With Ventoy, we can store multiple ISOs which it then presents as a boot screen.
|
||||
|
||||
To install Ventoy onto a flash drive, follow the [Get Started](https://www.ventoy.net/en/doc_start.html) guide on their website. Then once it's installed you should see a mountpoint labeled "Ventoy". There you can drop off any ISOs you want to appear on the boot screen. Here are some examples of what I keep in mine:
|
||||
|
||||
|
|
|
@ -46,7 +46,7 @@ Extend by 1G: `truncate -s +1G $HOME/scratch.img`
|
|||
|
||||
Shrink by 1G: `truncate -s -1G $HOME/scratch.img`
|
||||
|
||||
Check the filesystem to make sure that no inconsistencies occured. With `ext(2/3/4)` we can do this with the `e2fsck` command.
|
||||
Check the filesystem to make sure that no inconsistencies occurred. With `ext(2/3/4)` we can do this with the `e2fsck` command.
|
||||
|
||||
```bash
|
||||
e2fsck $HOME/scratch.img
|
||||
|
@ -81,4 +81,4 @@ rm $HOME/scratch.img
|
|||
|
||||
## Conclusion
|
||||
|
||||
With virtual disks we can experiment with different types of filesystems and perhaps try out snapshotting in supported filesystems. If we create virtual disks on [`tmpfs` ](/blog/lxdtmpfs/), then we can have a super fast file system as well!
|
||||
With virtual disks we can experiment with different types of filesystems and perhaps try out snapshoting in supported filesystems. If we create virtual disks on [`tmpfs` ](/blog/lxdtmpfs/), then we can have a super fast file system as well!
|
||||
|
|
|
@ -21,7 +21,7 @@ Of course you can replace the python version with whichever version you like. No
|
|||
source environment_name/bin/activate
|
||||
```
|
||||
|
||||
This now sets up your python interpretor and other utilities to use the installation in the `environment_name` folder. You can now install python packages using `pip` and have it only reside in this environment.
|
||||
This now sets up your python interpreter and other utilities to use the installation in the `environment_name` folder. You can now install python packages using `pip` and have it only reside in this environment.
|
||||
|
||||
To save all currently installed packages into `requirements.txt`:
|
||||
|
||||
|
|
|
@ -13,7 +13,7 @@ Implementing this protocol will allow one to interact with users on
|
|||
[Pixelfed](https://pixelfed.org/),
|
||||
[Friendica](https://friendi.ca/),
|
||||
and [others](http://fediverse.party/).
|
||||
Unfortuately for a static website, this protocol uses a
|
||||
Unfortunately for a static website, this protocol uses a
|
||||
publish-subscribe pattern. A service would need to be written
|
||||
in order to handle subscribers, pushing messages of published items,
|
||||
as well as receiving mentions.
|
||||
|
@ -32,7 +32,7 @@ lookup. For example if I am looking up the user `brozek@brandonrozek.com`,
|
|||
then the service will perform a GET request at
|
||||
`https://brandonrozek.com/.well-known/webfinger?resource=acct:brozek@brandonrozek.com`.
|
||||
|
||||
A webfinger request should return the handle of the user (the `subject`) as well as
|
||||
A Webfinger request should return the handle of the user (the `subject`) as well as
|
||||
a link to its `actor`.
|
||||
|
||||
```json
|
||||
|
@ -49,9 +49,9 @@ a link to its `actor`.
|
|||
}
|
||||
```
|
||||
|
||||
Now for a static site with one user, you can hardcode the webfinger directly
|
||||
Now for a static site with one user, you can hardcode the Webfinger directly
|
||||
at `.well-known/webfinger`. The only issue is that many webservers determine the
|
||||
mimetype of a file by its extension. In this case, we'll have to let the webserver
|
||||
mime-type of a file by its extension. In this case, we'll have to let the webserver
|
||||
(mine is nginx) know how to serve it.
|
||||
|
||||
```nginx
|
||||
|
|
|
@ -21,7 +21,7 @@ The main advantage of PESOS is that my primary interaction with Mastodon is thro
|
|||
- Boosts
|
||||
- Character Limits
|
||||
- Alt Description in Images
|
||||
- Post Permisssions
|
||||
- Post Permissions
|
||||
|
||||
## Disadvantages to PESOS
|
||||
|
||||
|
|
|
@ -7,7 +7,7 @@ tags: ["Networking"]
|
|||
|
||||
As of this time of writing, the current version of PiHole (5.0) supports adding custom DNS records, but not wildcard records. This makes it annoying if you run a bunch of different services within your LAN following a certain pattern.
|
||||
|
||||
Though since PiHole runs on top of `dnsmasq` it is easy to add an additional configuration file to point a domain containing `example.com` to a specific ip.
|
||||
Though since PiHole runs on top of `dnsmasq` it is easy to add an additional configuration file to point a domain containing `example.com` to a specific IP.
|
||||
|
||||
If you look in `/etc/dnsmasq.d/` there are the files `01-pihole.conf` and `02-pihole-dhcp.conf`. For our wildcard record, we're going to add a new file `03-custom-dns.conf`. Let's have an example where we want to map `example.com` and `*.example.com` to `192.168.0.10`.
|
||||
|
||||
|
|
|
@ -107,7 +107,7 @@ sudo ufw allow in on wg0 out on wg0
|
|||
|
||||
**On all machines:**
|
||||
|
||||
Have the wireguard service start at boot
|
||||
Have the Wireguard service start at boot
|
||||
|
||||
```bash
|
||||
sudo systemctl enable wg-quick@wg0
|
||||
|
|
|
@ -5,7 +5,7 @@ draft: false
|
|||
tags: [ "Linux" ]
|
||||
---
|
||||
|
||||
A dead simple way to send files between two linux machines not on the same network is to use a utility called [Magic Wormhole](https://github.com/warner/magic-wormhole). It is typically included in the standard repositories and is so simple the this blog post is going to end soon.
|
||||
A dead simple way to send files between two Linux machines not on the same network is to use a utility called [Magic Wormhole](https://github.com/warner/magic-wormhole). It is typically included in the standard repositories and is so simple the this blog post is going to end soon.
|
||||
|
||||
**Send a file:**
|
||||
|
||||
|
|
|
@ -32,7 +32,7 @@ Xephyr :$DNum \
|
|||
-resizeable &
|
||||
```
|
||||
|
||||
If we're going to be using this nested X server for one application, then I recommend using the `ratpoison` window manager as it will make the application full screen and supports the reizability of the Xephyr window.
|
||||
If we're going to be using this nested X server for one application, then I recommend using the `ratpoison` window manager as it will make the application full screen and supports the resizing of the Xephyr window.
|
||||
|
||||
```bash
|
||||
ratpoison -d :$DNum &
|
||||
|
|
Loading…
Reference in a new issue