Academy notes
Articles and updates from SoloQuant Academy.
Tag: nlp
Sentence Transformers v6.0 adds MultiVectorEncoder for ColBERT-style late interaction (MaxSim); absorbs PyLate/colpali paths; encode_query/encode_document; [Q] / [D] prompts. Breaking deps: transformers v5, torch≥2.2, huggingface-hub v1.x. Read the Migration Guide. Not investment advice.
Tetlock’s classic result: pessimism in a major WSJ column lines up with temporary downward pressure and higher trading volume—foundational alt-data intuition before modern transformers.
Yang, UY & Huang’s HKUST FinBERT is pretrained on financial communications—an alternative NLP backbone for filings and earnings text, distinct from Araci’s sentiment FinBERT.
Ke, Kelly, and Xiu (2019) show a supervised way to turn news text into return-prediction signals with careful screening. Plain-language intro for after-work Python NLP learners. Not investment advice.
Araci’s FinBERT fine-tunes BERT for financial sentiment so domain jargon scores more sensibly—then you can wire a Hugging Face checkpoint into a tiny daily-news feature pipeline.
Bybee, Kelly, Manela & Xiu turn decades of WSJ text into topic attention series that track the economy—a plain-language template for NLP alt-data practice in Python.
