Xiao-Yang Liu
@XiaoYangLiu10
Followers
172
Following
493
Media
0
Statuses
34
Ph.D. @Columbia U.
Manhattan, NY
Joined March 2020
We believe that the Hallucination issue is one major challenge before deploying FinGPT in real-world tasks. Here is an effort to reveal Hallucination behaviors. "Deficiency of Large Language Models in Finance: An Empirical Examination of Hallucination" https://t.co/cVgyxmnIec
0
2
12
#FinGPT available here
github.com
FinGPT: Open-Source Financial Large Language Models! Revolutionize 🔥 We release the trained model on HuggingFace. - AI4Finance-Foundation/FinGPT
@Columbia Ph.D. candidate @XiaoYangLiu10 shared his experiences leveraging the interplay between machine learning, signal processing, and computing for "Data-centric AI: From #ImageNet to Open-Source #FinRL and #FinGPT"
0
0
4
@Columbia Ph.D. candidate @XiaoYangLiu10 will share his experiences leveraging the interplay between machine learning, signal processing, and computing to achieve further progress. 🕓 May 3; 4:10 pm - 5:15 pm 📍 210 South Hall & Online Register now: https://t.co/VCNgB3qh5t
ischool.berkeley.edu
May 3, 2023, 4:10 pm - The creator of open-source projects FinRL, ElegantRL, and FinGPT outlines the deep learning revolution and his experiences applying it to the challenging domain of the financ...
0
3
3
2). Internet-scale finance data is critical, which should allow timely updates using an automatic data curation pipeline. #BloombergGPT has privileged data access and API access. A promising alternative is "democratizing Internet-scale finance data".
github.com
Democratizing Internet-scale financial data. Contribute to AI4Finance-Foundation/FinNLP development by creating an account on GitHub.
Why #FinGPT? Reasons: 1). Finance is high dynamic. #BloombergGPT retrains a LLM using a mixed dataset of finance and general data sources, which is too expensive (about 1.3M hours, at a cost $5M). Lightweight adaptation of #GPT4 is highly favorable.
1
1
5
3). Another key technology is "#RLHF (Reinforcement learning from human feedback)", which is missing in #BloombergGPT. RLHF enables learning individual preferences (risk-aversion level, investing habits, personalized robo-advisor, etc.)
2). Internet-scale finance data is critical, which should allow timely updates using an automatic data curation pipeline. #BloombergGPT has privileged data access and API access. A promising alternative is "democratizing Internet-scale finance data".
0
1
4
Why #FinGPT? Reasons: 1). Finance is high dynamic. #BloombergGPT retrains a LLM using a mixed dataset of finance and general data sources, which is too expensive (about 1.3M hours, at a cost $5M). Lightweight adaptation of #GPT4 is highly favorable.
github.com
FinGPT: Open-Source Financial Large Language Models! Revolutionize 🔥 We release the trained model on HuggingFace. - AI4Finance-Foundation/FinGPT
0
2
7
Let's rock it!
#FinGPT starts its journey! We borrow ideas from #ChatGPT, #GPT4, #BloombergGPT and stick to the #opensource and #openfinance culture. Welcome interested users to embrace this disruptive technology in the #AI4Finance interdisciplinary field! https://t.co/elrDvSRJGj
0
0
2
#ChatGPT for #FinTech. There are demos for #robo-advisor, #sentiment analysis, #alpha factors… https://t.co/kgXVGNN0QB
github.com
FinGPT: Open-Source Financial Large Language Models! Revolutionize 🔥 We release the trained model on HuggingFace. - AI4Finance-Foundation/FinGPT
0
2
6
Interesting video about using #chatGPT for algorithmic trading. It is quite interesting to use #reinforcementlearning to trade, and #FinRL is used as an example. Check it out
1
2
6
It was an interesting chat!
Meet with Yann Lecun and chat about financial big data as a challenging playground for AI/ML. Our NeurIPS project available at: https://t.co/eL2xlVQHjn
0
0
2
“A New Era of Massively Parallel Simulation: A Practical Tutorial Using ElegantRL” by Steven Li and Xiao-Yang Liu https://t.co/xNaLxcn0Ib
0
1
2
A tutorial of PaperTrading using Alpaca APIs is updated at FinRL-Meta. Now, an RL agent is automatically trading each day!
0
2
11
RL finds the optimal result of the MIMO beamforming task, which is known to be a nonconvex and NP-hard problem. Check the codes here
0
0
1
FinRL: The first open-source project for financial reinforcement learning. Please star. 🔥 - GitHub - AI4Finance-Foundation/FinRL: FinRL: The first open-source project for financial reinforcement learning. Please star. 🔥
0
2
5
“ElegantRL: Much Much More Stable than Stable-Baseline3” by Xiao-Yang Liu https://t.co/CZMMEwO1Hx
0
2
7
Hyperparameter Optimization using Ray tune for FinRL by @Athekunal
https://t.co/pPB2woXXoM
#art #AIart #machinelearning #deeplearning #MLsoGood #artificialintelligence #datascience #iiot #devops #data #code #python #bigdata #MLart #algorithm #DataScientist #Analytics #AI #VR
link.medium.com
In the previous articles on hyperparameter optimization using Optuna and Hyperparameter Sweep from Weights and Biases, we have gone…
0
8
4
A Medium blog:
medium.datadriveninvestor.com
Presented at NeurIPS Workshop on Data-Centric AI
FinRL-Meta: A Universe of Near-Real Market Environments for Data-Driven Deep Reinforcement Learning in Quantitative Finance. Liu, Rui, Gao, Yang, Yang, Wang, Wang, Guo: https://t.co/z1YX6UEiPk
1
0
5
FinRL for Quantitative Finance: plug-and-play DRL algorithms /by Bruce Yang https://t.co/qXtVb90NUc
#machinelearning #deeplearning #datascience #iiot #sztucznainteligencja #data #MLsoGood #code #python #bigdata #algorithm #programmer #pytorch #DataScientist #Analytics #AI #VR
link.medium.com
Compare several DRL libraries in one jupyter notebook
0
4
7
🔥🔥🔥 if you ever find reproducing/twisting existing deep reinforcement learning algorithms challenging (like we often did 😂), ElegantRL may be helpful : ) a lot of new features like distributed training are on the way, too https://t.co/NiMSWzP5BL
github.com
Massively Parallel Deep Reinforcement Learning. 🔥. Contribute to AI4Finance-Foundation/ElegantRL development by creating an account on GitHub.
ElegantRL - Lightweight, efficient and stable implementations of deep reinforcement learning algorithms using PyTorch.
0
3
32