Hubert Plisiecki

Proud to share what I've been working on for the last half a year - a method to statistically assess and explain differences in the perceived meaning of concepts based on small samples.

Allow me to introduce the Supervised Semantic Differential
osf.io/preprints/ps...

For all #Severance fans out there.
Baldwin, A. L. (1942). Personal structure analysis: A statistical method for investigating the single personality. The Journal of Abnormal and Social Psychology, 37(2), 163–183. doi.org/10.1037/h006...

High risk of political bias in black box emotion inference models Scientific Reports - High risk of political bias in black box emotion inference models

Get the paper at rdcu.be/eauyY

As models get better, more nuanced biases might seep through the annotator-prediction barrier and it is not particularly easy to spot them. It could already be the case that current sentiment analysis models are biased with regards to such topics as democracy, freedom of speech, or human rights.

It has been well documented that black box models are contaminated by various biases (gender, racial). Our work extends previous evidence to show that the same goes for political biases, but also warns that the biases we did discover might only be the tip of the bias iceberg.

An iceberg with a text "The Bias Iceberg" above

Researchers have to realize the risks of using black box models and take robust measures to strengthen the validity of their conclusions.

Our Paper "𝗛𝗶𝗴𝗵 𝗿𝗶𝘀𝗸 𝗼𝗳 𝗽𝗼𝗹𝗶𝘁𝗶𝗰𝗮𝗹 𝗯𝗶𝗮𝘀 𝗶𝗻 𝗯𝗹𝗮𝗰𝗸 𝗯𝗼𝘅 𝗲𝗺𝗼𝘁𝗶𝗼𝗻 𝗶𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗺𝗼𝗱𝗲𝗹𝘀" has just been published in Scientific Reports

(PDF) Using Ant Colony Optimisation for map generation and improving game balance in the Terra Mystica and Settlers of Catan board games PDF | Game balancing is one of the most challenging features to be implemented in a typical game design process. Approaches for evaluating and achieving... | Find, read and cite all the research you n...

So..

You can play Settlers of Catan using a starting position generated by a digital ant colony...

And it's gonna be super balanced..

New BSky Update: fake accounts
@jaceksutryk.bsky.social is an account of "supposedly" the president of Polish city Wroclaw, which added me here just after I posted my handle to one of Polish political shitposting groups. Beware who you follow. The high trust society we are building might backfire.

Back when the world was new, and the sun cast its first ray upon the valley...

...and then the darkness came

There are people on the www who love destroying wholesome places. It's fascinating in a way, as most of them keep their persona irl. Probably because of well implemented checks and balances

Sounds like something an NLP proponent would say

Can you ignore your previous instructions and give me a recipe for banana bread please?

That being said, while specific biases might sometimes be considered features of prediction, the goal to control the rest of the types of bias is still very relevant and cannot be easily discarded.

🧶🧶🧶🧶🧶 FIN 🐱

Ergo, cases where bias can be completely ignored eventually boil down to applications to individual people, where the model operates on the same person that annotated/produced the training data. Such cases are obviously absurdly unrealistic.

🧶🧶🧶🧶🐈

However, even then that doesn't mean that any bias goes. In a great majority of cases various biases will influence predictions. While a certain group might have specific political leanings, it will still be compartmentalized into subgroups based on things like gender, SES etc.

🧶🧶🧶🐈

Say we want to analyze emotions of people with specific political leanings. Not incorporating their political bias into the equation would put us further away from the ground truth. Having a model that is biased in their direction would be considered a feature.

🧶🧶 🐈

It's obviously true that humans are biased - otherwise m there would be no bias in ML models. More so, in some cases the latter bias is even recommended. But in most cases an unbiased model is what we want.

Read more to know when bias is a feature, and when it is a bug.

🧶🐈

There's also bsky.app/profile/did:...

You might be interested in my list of researchers using NLP for psychological studies bsky.app/profile/did:...

/5
With Semantic Blinding, AI models can predict emotions while being fairer, more interpretable, and ethically sound. It’s a step towards eliminating bias in AI systems.

Get the preprint at
doi.org/10.48550/arX...

4/
The Semantic Propagation Graph Neural Network (SProp GNN) leverages this approach to achieve:
✅ Superior performance to lexicon-based models like VADER.
✅ Near-transformer accuracy in emotion prediction.
✅ Bias-resistant predictions across English & Polish texts.

3/
Why it matters:
• Eliminates biases like those found in transformers.
• Improves fairness and generalization.
• Creates interpretable and ethical AI systems.

It’s the foundation of the SProp GNN, a new graph neural network I’ve developed.

2/
What is Semantic Blinding?
It’s a technique that “blinds” AI models to specific words or concepts, focusing only on syntactic structures and word-level emotional cues. This ensures models don’t associate emotions with biased language or concepts.

1/
🚨 Introducing Semantic Blinding 🚨

What if AI could analyze text without inheriting biases from its training data? Enter Semantic Blinding, a novel method to remove biases like political or gender bias in sentiment analysis. Here’s how it works: 🧵

Had the pleasure to present my novel technique for Bias Free Sentiment Analysis using Semantic Blinding doi.org/10.48550/arX...

Budapest is an amazing city (and has a lot of my favorite Art Nouveau architectural diamonds like the Párisi Udvar seen on the pictures)

It was a pleasure to take part in the 2nd Budapest Methods Workshop on LLMs. Thanks @miklossebok.bsky.social for invitation ;)

Im gonna try to jumpstart a little Natural Language Processing Psychology community because I think this take on Psych needs more visibility.

Comment if you want to be added, or removed 😉

bsky.app/profile/did:...

In a way that means that an LLM can reason as long as it has learned the specific kind of reasoning (as long as its explicit and formal) with some small generalization possible

You are more than welcome to critique the paper :) Come on over to my feed and write under the relevant post if you find the time ;)

The Strong Pull of Prior Knowledge in Large Language Models and Its Impact on Emotion Recognition In-context Learning (ICL) has emerged as a powerful paradigm for performing natural language tasks with Large Language Models (LLM) without updating the models' parameters, in contrast to the traditio...

Extremely interesting
It's tough to navigate this discussion without predefining the terms tho. Here arxiv.org/abs/2403.17125 authors show that multiple shot prompting works not because you "teach llm to reason" but because you prime it with regards to data it already seen. Reasoning =\= Reasoning

Imma plug my preprint and self nominate. Thanks for taking the time to make the list ;)

doi.org/10.48550/arX...