Scaling Agent Learning via Experience Synthesis
Paper
β’
2511.03773
β’
Published
β’
80
AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
Don't Waste It: Guiding Generative Recommenders with Structured Human Priors via Multi-head Decoding
Scaling Agent Learning via Experience Synthesis
Benjamin-eecsΒ
authored
2
papers
about 1 month ago
SelmaNajih001Β
posted
an
update
about 1 month ago
Post
2831
How Financial News Can Be Used to Train Good Financial Models π°
Numbers tell you what happened, but news tells you why.
Iβve written an article explaining how news can be used to train AI models for sentiment analysis and better forecasting. Hope you find it interesting!
Read it here: https://huggingface.co/blog/SelmaNajih001/llms-applied-to-finance
I would love to read your opinions! Iβm open to suggestions on how to improve the methodology and the training
Numbers tell you what happened, but news tells you why.
Iβve written an article explaining how news can be used to train AI models for sentiment analysis and better forecasting. Hope you find it interesting!
Read it here: https://huggingface.co/blog/SelmaNajih001/llms-applied-to-finance
I would love to read your opinions! Iβm open to suggestions on how to improve the methodology and the training
- 1 reply
SelmaNajih001Β
posted
an
update
about 1 month ago
Post
3022
Which is the best model to use as a signal for investment?
Here who is gaining the most:
SelmaNajih001/InvestmentStrategyBasedOnSentiment
The Space uses titles from this dataset:
π SelmaNajih001/Cnbc_MultiCompany
Given a news title, it calculates a sentiment score : if the score crosses a certain threshold, the strategy decides to buy or sell.
Each trade lasts one day, and the strategy then computes the daily return.
For Tesla the best model seems to be the regression π
Just a quick note: the model uses the closing price as the buy price, meaning it already reflects the impact of the news.
Here who is gaining the most:
SelmaNajih001/InvestmentStrategyBasedOnSentiment
The Space uses titles from this dataset:
π SelmaNajih001/Cnbc_MultiCompany
Given a news title, it calculates a sentiment score : if the score crosses a certain threshold, the strategy decides to buy or sell.
Each trade lasts one day, and the strategy then computes the daily return.
For Tesla the best model seems to be the regression π
Just a quick note: the model uses the closing price as the buy price, meaning it already reflects the impact of the news.
SelmaNajih001Β
posted
an
update
about 2 months ago
Post
679
How Financial News Can Be Used to Train Good Financial Models π°
Numbers tell you what happened, but news tells you why.
Iβve written an article explaining how news can be used to train AI models for sentiment analysis and better forecasting. Hope you find it interesting!
Read it here: https://huggingface.co/blog/SelmaNajih001/llms-applied-to-finance
I would love to read your opinions! Iβm open to suggestions on how to improve the methodology and the training
Numbers tell you what happened, but news tells you why.
Iβve written an article explaining how news can be used to train AI models for sentiment analysis and better forecasting. Hope you find it interesting!
Read it here: https://huggingface.co/blog/SelmaNajih001/llms-applied-to-finance
I would love to read your opinions! Iβm open to suggestions on how to improve the methodology and the training
- 1 reply
Benjamin-eecsΒ
authored
a
paper
about 2 months ago
SelmaNajih001Β
posted
an
update
about 2 months ago
Post
385
Which is the best model to use as a signal for investment? π€
Iβve created a Space where you can compare three models:
-Two available on my profile
- ProsusAI/finbert
You can try it here:
π SelmaNajih001/InvestmentStrategyBasedOnSentiment
The Space uses titles from this dataset:
π SelmaNajih001/Cnbc_MultiCompany
Given a news title, it calculates a sentiment score : if the score crosses a certain threshold, the strategy decides to buy or sell.
Each trade lasts one day, and the strategy then computes the daily return.
Just a quick note: the model uses the closing price as the buy price, meaning it already reflects the impact of the news.
If I had chosen the opening price, the results would have been less biased but less realistic given the data available.
Iβve created a Space where you can compare three models:
-Two available on my profile
- ProsusAI/finbert
You can try it here:
π SelmaNajih001/InvestmentStrategyBasedOnSentiment
The Space uses titles from this dataset:
π SelmaNajih001/Cnbc_MultiCompany
Given a news title, it calculates a sentiment score : if the score crosses a certain threshold, the strategy decides to buy or sell.
Each trade lasts one day, and the strategy then computes the daily return.
Just a quick note: the model uses the closing price as the buy price, meaning it already reflects the impact of the news.
If I had chosen the opening price, the results would have been less biased but less realistic given the data available.
LeitianΒ
authored
a
paper
about 2 months ago
Benjamin-eecsΒ
authored
a
paper
about 2 months ago
LeitianΒ
authored
a
paper
about 2 months ago
SelmaNajih001Β
posted
an
update
2 months ago
Post
2506
I found it excellent and very well done.
One of the best explanations of embedding I've ever read. Well done, @hesamation !
Had to share this: hesamation/primer-llm-embedding
One of the best explanations of embedding I've ever read. Well done, @hesamation !
Had to share this: hesamation/primer-llm-embedding
SelmaNajih001Β
posted
an
update
2 months ago
Post
2286
Finally, I uploaded the model I developed for my masterβs thesis! Given a financial event, it provides explained predictions based on a dataset of past news and central bank speeches.
Try it out here:
SelmaNajih001/StockPredictionExplanation
(Just restart the space and wait a minute)
The dataset used for RAG can be found here:
SelmaNajih001/FinancialNewsAndCentralBanksSpeeches-Summary-Rag
While the dataset used for the training is:
SelmaNajih001/FinancialClassification
I also wrote an article to explain how I've done the training. You can find it here https://huggingface.co/blog/SelmaNajih001/explainable-financial-predictions
Try it out here:
SelmaNajih001/StockPredictionExplanation
(Just restart the space and wait a minute)
The dataset used for RAG can be found here:
SelmaNajih001/FinancialNewsAndCentralBanksSpeeches-Summary-Rag
While the dataset used for the training is:
SelmaNajih001/FinancialClassification
I also wrote an article to explain how I've done the training. You can find it here https://huggingface.co/blog/SelmaNajih001/explainable-financial-predictions
-
- 2 replies
SelmaNajih001Β
posted
an
update
2 months ago
Post
3391
Introducing a Hugging Face Tutorial on Regression
While Hugging Face offers extensive tutorials on classification and NLP tasks, there is very little guidance on performing regression tasks with Transformers.
In my latest article, I provide a step-by-step guide to running regression using Hugging Face, applying it to financial news data to predict stock returns.
In this tutorial, you will learn how to:
-Prepare and preprocess textual and numerical data for regression
-Configure a Transformer model for regression tasks
-Apply the model to real-world financial datasets with fully reproducible code
Read the full article here: https://huggingface.co/blog/SelmaNajih001/how-to-run-a-regression-using-hugging-face
The dataset used: SelmaNajih001/FinancialClassification
While Hugging Face offers extensive tutorials on classification and NLP tasks, there is very little guidance on performing regression tasks with Transformers.
In my latest article, I provide a step-by-step guide to running regression using Hugging Face, applying it to financial news data to predict stock returns.
In this tutorial, you will learn how to:
-Prepare and preprocess textual and numerical data for regression
-Configure a Transformer model for regression tasks
-Apply the model to real-world financial datasets with fully reproducible code
Read the full article here: https://huggingface.co/blog/SelmaNajih001/how-to-run-a-regression-using-hugging-face
The dataset used: SelmaNajih001/FinancialClassification
- 1 reply
Benjamin-eecsΒ
authored
3
papers
2 months ago
SelmaNajih001Β
posted
an
update
2 months ago
Post
1654
Introducing
SelmaNajih001/StockPredictionExplanation, built with GRPO and RAG:
-GRPO trains the model to predict and explain stock direction.
-RAG grounds explanations in historical financial news and central bank speeches.
Together, they create a system that forecasts stock movements and shows the reasoning behind them.
Full article: Explainable Financial Predictions β https://huggingface.co/blog/SelmaNajih001/explainable-financial-predictions
Try it here: StockPredictionExplanation Space β SelmaNajih001/StockPredictionExplanation
-GRPO trains the model to predict and explain stock direction.
-RAG grounds explanations in historical financial news and central bank speeches.
Together, they create a system that forecasts stock movements and shows the reasoning behind them.
Full article: Explainable Financial Predictions β https://huggingface.co/blog/SelmaNajih001/explainable-financial-predictions
Try it here: StockPredictionExplanation Space β SelmaNajih001/StockPredictionExplanation
SelmaNajih001Β
posted
an
update
2 months ago
Post
265
Predicting Stock Price Movements from News π°π
I trained a model to predict stock price movements (Up, Down, Neutral) from company news.
Dataset: Articles linked to next-day price changes, covering Apple, Tesla, and more.
Approach: Fine-tuned allenai/longformer-base-4096 for classification.
Outcome: The model captures the link between news and stock movements, handling long articles and producing probability scores for each label.
Comparison: Shows promising alignment with stock trends, sometimes outperforming FinBERT.
Feel free to try the model and explore how news can influence stock predictions SelmaNajih001/SentimentAnalysis
I trained a model to predict stock price movements (Up, Down, Neutral) from company news.
Dataset: Articles linked to next-day price changes, covering Apple, Tesla, and more.
Approach: Fine-tuned allenai/longformer-base-4096 for classification.
Outcome: The model captures the link between news and stock movements, handling long articles and producing probability scores for each label.
Comparison: Shows promising alignment with stock trends, sometimes outperforming FinBERT.
Feel free to try the model and explore how news can influence stock predictions SelmaNajih001/SentimentAnalysis
AnSungJae3489Β
posted
an
update
2 months ago
Post
2600
ShareGPT? How about ShareGPT-X?
We release **92K** Human with LLM conversations as a refresh and update over the original ShareGPT Dataset.
DSULT-Core/ShareGPT-X
We release **92K** Human with LLM conversations as a refresh and update over the original ShareGPT Dataset.
DSULT-Core/ShareGPT-X
Benjamin-eecsΒ
authored
a
paper
3 months ago