#classifying — Public Fediverse posts

Classifying aviation-related posts on Hacker News with SLMs

https://www.skysight.inc/blog/hacker-news-aviation

#HackerNews #Classifying #aviation-related #posts #on #Hacker #News #with #SLMs #aviation #machinelearning #hackernews #SLMs #technology

Classifying aviation-related posts on Hacker News with SLMs

https://www.skysight.inc/blog/hacker-news-aviation

#HackerNews #Classifying #aviation-related #posts #on #Hacker #News #with #SLMs #aviation #machinelearning #hackernews #SLMs #technology

Classifying aviation-related posts on Hacker News with SLMs

https://www.skysight.inc/blog/hacker-news-aviation

#HackerNews #Classifying #aviation-related #posts #on #Hacker #News #with #SLMs #aviation #machinelearning #hackernews #SLMs #technology

Classifying aviation-related posts on Hacker News with SLMs

https://www.skysight.inc/blog/hacker-news-aviation

#HackerNews #Classifying #aviation-related #posts #on #Hacker #News #with #SLMs #aviation #machinelearning #hackernews #SLMs #technology

(19/N) Let's now turn to the third question of the #ThreatModelingManifesto:

3. What are you going to do about it?

It pays to first establish a few contraints for what you can do, in theory, by #classifying your #assets. Again, for an individual human being, opposed to organizations or companies, it's nearly impossible to impose principles like #ZeroTrust or #NeedToKnow on personal relationships, the closer they get.

So, avoid recycling terms from popular, but less intuitive schemes: Fanciful intelligence labels like “top secret”, “confidential”, or “unclassified” do not tell you what goes into the respective box, and how to handle access to it.

Add another column to your assets spreadsheet, label it "Classification", and pick a more human-centered approach for its values, like:

• For Your Eyes Only (FYEO)
• Intimate
• None Of Your Business (NOYB)
• Shared
• Public

Let's briefly go through these suggestions:

For Your Eyes Only (FYEO)

Assets that are only accessible to, and controlled by nobody but you, because they need to be resilient, even in the face of the closest of your close people misbehaving. Preferably, these assets are kept publicly undetectable and unknown. When you are gone, these assets will be gone, too. FYEO does not make a good default class, though.

Start of this thread:
https://mastodon.de/@tuxwise/113503228291818865

#ThreatModeling #4D

(19/N) Let's now turn to the third question of the #ThreatModelingManifesto:

3. What are you going to do about it?

It pays to first establish a few contraints for what you can do, in theory, by #classifying your #assets. Again, for an individual human being, opposed to organizations or companies, it's nearly impossible to impose principles like #ZeroTrust or Need-to-know on personal relationships, the closer they get.

So, avoid recycling terms from popular, but less intuitive schemes: Fanciful intelligence labels like “top secret”, “confidential”, or “unclassified” do not tell you what goes into the respective box, and how to handle access to it.

Add another column to your assets spreadsheet, label it "Classification", and pick a more human-centered approach for its values, like:

• 🤐 For Your Eyes Only (FYEO)
• 🔒 Intimate
• 🚪 None Of Your Business (NOYB)
• 🤝 Shared
• 🫴 Public

Let's briefly go through these suggestions:

🤐 For Your Eyes Only (FYEO)

Assets that are only accessible to, and controlled by nobody but you, because they need to be resilient, even in the face of the closest of your close people misbehaving. Preferably, these assets are kept publicly undetectable and unknown. When you are gone, these assets will be gone, too. FYEO does not make a good default class, though.

Start of this thread:
https://mastodon.de/@tuxwise/113503228291818865

#ThreatModeling #4D

(19/N) Let's now turn to the third question of the #ThreatModelingManifesto:

3. What are you going to do about it?

It pays to first establish a few contraints for what you can do, in theory, by #classifying your #assets. Again, for an individual human being, opposed to organizations or companies, it's nearly impossible to impose principles like #ZeroTrust or #NeedToKnow on personal relationships, the closer they get.

So, avoid recycling terms from popular, but less intuitive schemes: Fanciful intelligence labels like “top secret”, “confidential”, or “unclassified” do not tell you what goes into the respective box, and how to handle access to it.

Add another column to your assets spreadsheet, label it "Classification", and pick a more human-centered approach for its values, like:

• For Your Eyes Only (FYEO)
• Intimate
• None Of Your Business (NOYB)
• Shared
• Public

Let's briefly go through these suggestions:

For Your Eyes Only (FYEO)

Assets that are only accessible to, and controlled by nobody but you, because they need to be resilient, even in the face of the closest of your close people misbehaving. Preferably, these assets are kept publicly undetectable and unknown. When you are gone, these assets will be gone, too. FYEO does not make a good default class, though.

Start of this thread:
https://mastodon.de/@tuxwise/113503228291818865

#ThreatModeling #4D

(19/N) Let's now turn to the third question of the #ThreatModelingManifesto:

3. What are you going to do about it?

It pays to first establish a few contraints for what you can do, in theory, by #classifying your #assets. Again, for an individual human being, opposed to organizations or companies, it's nearly impossible to impose principles like #ZeroTrust or #NeedToKnow on personal relationships, the closer they get.

So, avoid recycling terms from popular, but less intuitive schemes: Fanciful intelligence labels like “top secret”, “confidential”, or “unclassified” do not tell you what goes into the respective box, and how to handle access to it.

Add another column to your assets spreadsheet, label it "Classification", and pick a more human-centered approach for its values, like:

• For Your Eyes Only (FYEO)
• Intimate
• None Of Your Business (NOYB)
• Shared
• Public

Let's briefly go through these suggestions:

For Your Eyes Only (FYEO)

Assets that are only accessible to, and controlled by nobody but you, because they need to be resilient, even in the face of the closest of your close people misbehaving. Preferably, these assets are kept publicly undetectable and unknown. When you are gone, these assets will be gone, too. FYEO does not make a good default class, though.

Start of this thread:
https://mastodon.de/@tuxwise/113503228291818865

#ThreatModeling #4D

(19/N) Let's now turn to the third question of the #ThreatModelingManifesto:

3. What are you going to do about it?

It pays to first establish a few contraints for what you can do, in theory, by #classifying your #assets. Again, for an individual human being, opposed to organizations or companies, it's nearly impossible to impose principles like #ZeroTrust or #NeedToKnow on personal relationships, the closer they get.

So, avoid recycling terms from popular, but less intuitive schemes: Fanciful intelligence labels like “top secret”, “confidential”, or “unclassified” do not tell you what goes into the respective box, and how to handle access to it.

Add another column to your assets spreadsheet, label it "Classification", and pick a more human-centered approach for its values, like:

• For Your Eyes Only (FYEO)
• Intimate
• None Of Your Business (NOYB)
• Shared
• Public

Let's briefly go through these suggestions:

For Your Eyes Only (FYEO)

Assets that are only accessible to, and controlled by nobody but you, because they need to be resilient, even in the face of the closest of your close people misbehaving. Preferably, these assets are kept publicly undetectable and unknown. When you are gone, these assets will be gone, too. FYEO does not make a good default class, though.

Start of this thread:
https://mastodon.de/@tuxwise/113503228291818865

#ThreatModeling #4D

@cnDuKeli @tcdh

The approach they used for #classifying the poems' function in the narrative were interesting, using #LLMs. Performance is pretty bad so far. Keli took up the call for #openness about #failure from a session this morning and showed that the different models are bad in different ways, which allowed the team (and us) to learn something about the models. I think that's great and valuable!

#ErrorAnalysis #ConfusionMatrix #MajorityBaseline #DH2024

'Principled Out-of-Distribution Detection via Multiple Testing', by Akshayaa Magesh, Venugopal V. Veeravalli, Anirban Roy, Susmit Jha.

http://jmlr.org/papers/v24/23-0838.html

#detecting #distribution #classifying

I have plenty more achievable goals for https://schizo.social (like multi-account, or #Calckey support) but something I'd love to try is #classifying posts with #machineLearning #tfidf

I'd like to be able to define "labels" and then train it to identify those on the fly. Then either mute or highlight posts that #classify highly.

Not so much an #algorithm, as a #filter.

#ai #ml #webDev #classifier

I have plenty more achievable goals for https://schizo.social (like multi-account, or #Calckey support) but something I'd love to try is #classifying posts with #machineLearning #tfidf

I'd like to be able to define "labels" and then train it to identify those on the fly. Then either mute or highlight posts that #classify highly.

Not so much an #algorithm, as a #filter.

#ai #ml #webDev #classifier

I have plenty more achievable goals for https://schizo.social (like multi-account, or #Calckey support) but something I'd love to try is #classifying posts with #machineLearning #tfidf

I'd like to be able to define "labels" and then train it to identify those on the fly. Then either mute or highlight posts that #classify highly.

Not so much an #algorithm, as a #filter.

#ai #ml #webDev #classifier

I have plenty more achievable goals for https://schizo.social (like multi-account, or #Calckey support) but something I'd love to try is #classifying posts with #machineLearning #tfidf

I'd like to be able to define "labels" and then train it to identify those on the fly. Then either mute or highlight posts that #classify highly.

Not so much an #algorithm, as a #filter.

#ai #ml #webDev #classifier

I have plenty more achievable goals for https://schizo.social (like multi-account, or #Calckey support) but something I'd love to try is #classifying posts with #machineLearning #tfidf

I'd like to be able to define "labels" and then train it to identify those on the fly. Then either mute or highlight posts that #classify highly.

Not so much an #algorithm, as a #filter.

#ai #ml #webDev #classifier

A Reproducible and Realistic Evaluation of Partial Domain Adaptation Methods

Tiago Salvador, Kilian FATRAS, Ioannis Mitliagkas, Adam M Oberman

Action editor: Mingsheng Long.

https://openreview.net/forum?id=XcVzIBXeRn

#adaptation #classifying #leveraging

A Reproducible and Realistic Evaluation of Partial Domain Adaptation Methods

Tiago Salvador, Kilian FATRAS, Ioannis Mitliagkas, Adam M Oberman

Action editor: Mingsheng Long.

https://openreview.net/forum?id=XcVzIBXeRn

#adaptation #classifying #leveraging

A Reproducible and Realistic Evaluation of Partial Domain Adaptation Methods

Tiago Salvador, Kilian FATRAS, Ioannis Mitliagkas, Adam M Oberman

Action editor: Mingsheng Long.

https://openreview.net/forum?id=XcVzIBXeRn

#adaptation #classifying #leveraging

A Reproducible and Realistic Evaluation of Partial Domain Adaptation Methods

Tiago Salvador, Kilian FATRAS, Ioannis Mitliagkas, Adam M Oberman

Action editor: Mingsheng Long.

https://openreview.net/forum?id=XcVzIBXeRn

#adaptation #classifying #leveraging

A Reproducible and Realistic Evaluation of Partial Domain Adaptation Methods

Tiago Salvador, Kilian FATRAS, Ioannis Mitliagkas, Adam M Oberman

Action editor: Mingsheng Long.

https://openreview.net/forum?id=XcVzIBXeRn

#adaptation #classifying #leveraging