arxiv: 1908.10063 · v1 · submitted 2019-08-27 · 💻 cs.CL · cs.LG

Recognition: 2 theorem links

· Lean Theorem

FinBERT: Financial Sentiment Analysis with Pre-trained Language Models

Dogu Araci

Authors on Pith no claims yet

Pith reviewed 2026-05-15 20:22 UTC · model grok-4.3

classification 💻 cs.CL cs.LG

keywords FinBERTfinancial sentiment analysisBERTpre-trained language modelsdomain adaptationsentiment classificationfinancial NLP

0 comments

The pith

FinBERT adapts a BERT model with financial text to improve sentiment classification on specialized datasets.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces FinBERT to handle financial sentiment analysis, where specialized language and scarce labeled data limit general-purpose models. It tests the idea that further pre-training BERT on financial corpora, followed by partial fine-tuning, yields stronger results than existing approaches. The work shows gains across metrics on two datasets even when using smaller training sets. A reader would care because this points to a practical way to build better tools for analyzing financial text without needing huge amounts of new labeled examples.

Core claim

FinBERT is a language model based on BERT that receives additional pre-training on financial corpora. When applied to financial sentiment analysis, even with a smaller training set and fine-tuning only part of the model, it improves every measured metric over current state-of-the-art results on two financial sentiment datasets and outperforms traditional machine learning methods.

What carries the argument

FinBERT, a BERT model further pre-trained on financial text and then partially fine-tuned for sentiment classification tasks.

If this is right

Sentiment classification reaches higher accuracy on standard financial benchmarks than earlier methods.
Effective results appear even when the amount of labeled financial training data is reduced.
Partial fine-tuning of the model is enough to realize the performance gains.
Domain-adapted pre-trained models outperform both general language models and conventional machine learning baselines in this setting.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same pre-training plus partial fine-tuning pattern could transfer to other specialized domains such as legal or medical text analysis.
It suggests that large pre-trained models can be adapted for narrow tasks without retraining every parameter from scratch.
Real-time monitoring of financial news for sentiment signals becomes more feasible with smaller labeled datasets.

Load-bearing premise

That pre-training on financial corpora will create representations that transfer to better sentiment classification than general models when labeled data remains limited.

What would settle it

Testing FinBERT on a new financial sentiment dataset and finding it fails to exceed the prior best model on accuracy, F1, or related metrics would disprove the central performance claim.

read the original abstract

Financial sentiment analysis is a challenging task due to the specialized language and lack of labeled data in that domain. General-purpose models are not effective enough because of the specialized language used in a financial context. We hypothesize that pre-trained language models can help with this problem because they require fewer labeled examples and they can be further trained on domain-specific corpora. We introduce FinBERT, a language model based on BERT, to tackle NLP tasks in the financial domain. Our results show improvement in every measured metric on current state-of-the-art results for two financial sentiment analysis datasets. We find that even with a smaller training set and fine-tuning only a part of the model, FinBERT outperforms state-of-the-art machine learning methods.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

FinBERT applies BERT to financial sentiment via extra domain pre-training and reports gains on two datasets, but the lack of ablations leaves the source of those gains unclear.

read the letter

FinBERT applies the BERT model to financial sentiment analysis by doing additional pre-training on domain-specific text. It claims better performance than prior state-of-the-art on two datasets, even with smaller training sets and only partial fine-tuning. This is a direct and useful extension of the BERT work into a new domain where language is specialized and labeled data is scarce. The paper correctly notes that general-purpose models struggle here and that pre-trained language models can reduce the need for lots of labels. The experimental setup follows standard transfer learning practices, which is appropriate for the problem. What stands out is the focus on practical constraints like limited labeled data, which matches real-world financial applications in market monitoring and risk systems. If the gains hold up, this gives practitioners a stronger starting point for tasks like analyzing news or reports for sentiment. The soft spots are in the validation of the claims. The abstract gives no specific numbers or details on the baselines used, and there are no ablations to isolate whether the financial pre-training or the partial fine-tuning is responsible for the improvements. This leaves room for the results to be explained by other factors such as hyperparameter tuning or particular data splits. The assumption that this combination works reliably with limited data needs stronger support through controlled experiments. This paper is for researchers and engineers working on financial NLP tasks who want a domain-adapted model or a template for similar adaptations. A reader interested in applying pre-trained models to specialized text would find the description helpful as a case study. I would recommend sending it for peer review. The core idea is clear and the motivation is solid, but the authors should add the missing ablations and full experimental details to make the performance claims more robust.

Referee Report

2 major / 2 minor

Summary. The paper introduces FinBERT, a BERT-based language model that undergoes additional pre-training on financial-domain corpora. It evaluates the model on two financial sentiment analysis datasets and claims consistent improvements over prior state-of-the-art results in all reported metrics, even when using a smaller training set and fine-tuning only a subset of the parameters.

Significance. If the empirical gains are shown to be robust and attributable to the domain-specific pre-training step, the work would supply concrete evidence that continued pre-training on specialized text can mitigate data scarcity in domain-specific NLP tasks, extending transfer-learning benefits to financial applications.

major comments (2)

[Experiments] Experiments section: the headline claim that domain pre-training plus partial fine-tuning reliably outperforms both general-purpose models and traditional ML baselines requires ablations that isolate each component (e.g., vanilla BERT under the same partial-fine-tuning regime, full-model fine-tuning of FinBERT, and identical hyper-parameter search); without these controls the numerical improvements cannot be confidently attributed to the proposed mechanisms rather than tuning or split variation.
[Results] Results tables / abstract: no numerical metric values, baseline descriptions, dataset sizes, train/test splits, or statistical significance tests are supplied, so the assertion of improvement “in every measured metric” on current SOTA cannot be verified from the manuscript.

minor comments (2)

[Abstract] Abstract: the phrase “current state-of-the-art results” should name the specific prior methods and papers being compared.
[Introduction] Introduction: specify the exact training-set sizes for the two datasets and how they relate to the sizes used in the cited SOTA baselines.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We address each major comment below and will revise the manuscript to strengthen the experimental design and clarity of results.

read point-by-point responses

Referee: [Experiments] Experiments section: the headline claim that domain pre-training plus partial fine-tuning reliably outperforms both general-purpose models and traditional ML baselines requires ablations that isolate each component (e.g., vanilla BERT under the same partial-fine-tuning regime, full-model fine-tuning of FinBERT, and identical hyper-parameter search); without these controls the numerical improvements cannot be confidently attributed to the proposed mechanisms rather than tuning or split variation.

Authors: We agree that the current set of experiments does not fully isolate the contributions of domain-specific pre-training versus partial fine-tuning. In the revised version we will add the requested ablations: (1) vanilla BERT under the identical partial-fine-tuning regime, (2) full-model fine-tuning of FinBERT, and (3) a common hyper-parameter search budget applied to all models. These additions will allow readers to attribute performance differences more confidently to the domain-adaptation step. revision: yes
Referee: [Results] Results tables / abstract: no numerical metric values, baseline descriptions, dataset sizes, train/test splits, or statistical significance tests are supplied, so the assertion of improvement “in every measured metric” on current SOTA cannot be verified from the manuscript.

Authors: We acknowledge that the manuscript version reviewed does not present the concrete numerical values, baseline details, split sizes, or significance tests in a readily verifiable form. We will expand the results section and tables to report all metric values explicitly, describe each baseline, state dataset sizes and train/test splits, and include statistical significance tests (e.g., paired t-tests or McNemar’s test) so that the claim of consistent improvement can be directly verified. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical transfer-learning experiment with no derivations

full rationale

The paper is a standard empirical study: it further pre-trains BERT on financial corpora and fine-tunes for sentiment classification, then reports accuracy/F1 gains on two labeled datasets. No equations, no derivations, no fitted parameters renamed as predictions, and no self-citation chains that bear the central claim. All performance numbers are external measurements against baselines; nothing reduces to its own inputs by construction. This is the expected non-circular outcome for a pure experimental transfer-learning paper.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract-only review supplies no explicit free parameters, axioms, or invented entities; the central claim rests on the unstated assumption that standard BERT fine-tuning transfers effectively to finance.

axioms (1)

domain assumption Pre-trained language models require fewer labeled examples for effective domain adaptation
Explicit hypothesis stated in the abstract.

pith-pipeline@v0.9.0 · 5405 in / 1184 out tokens · 30429 ms · 2026-05-15T20:22:30.685641+00:00 · methodology

discussion (0)

Forward citations

Cited by 18 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

PolyBench: Benchmarking LLM Forecasting and Trading Capabilities on Live Prediction Market Data
q-fin.CP 2026-04 conditional novelty 8.0

Only two of seven LLMs produce positive returns on live Polymarket data, with MiMo-V2-Flash at 17.6% CWR and Gemini-3-Flash at 6.2% CWR while the other five lose money.
VertMark: A Unified Training-Free Robust Watermarking Framework for Vertical Domain Pre-trained Language Models
cs.CR 2026-05 unverdicted novelty 7.0

VertMark embeds robust, training-free watermarks into vertical domain language models by creating hidden semantic equivalence between low-frequency triggers and high-frequency domain terms via parameter swaps, support...
AgentPulse: A Continuous Multi-Signal Framework for Evaluating AI Agents in Deployment
cs.AI 2026-04 conditional novelty 7.0

AgentPulse is a continuous multi-signal framework that scores AI agents on benchmark performance, adoption, sentiment and ecosystem health, showing these factors are complementary and that benchmark-plus-sentiment pre...
Rethinking Scale: Deployment Trade-offs of Small Language Models under Agent Paradigms
cs.CL 2026-04 unverdicted novelty 7.0

Single-agent systems with tools provide the optimal performance-efficiency trade-off for small language models, outperforming base models and multi-agent setups.
Agentic Retrieval-Augmented Generation for Financial Document Question Answering
cs.AI 2026-05 unverdicted novelty 6.0

FinAgent-RAG achieves 76.81-78.46% execution accuracy on financial QA benchmarks by combining contrastive retrieval, program-of-thought code generation, and adaptive strategy routing, outperforming baselines by 5.62-9...
Temporal Reasoning Is Not the Bottleneck: A Probabilistic Inconsistency Framework for Neuro-Symbolic QA
cs.AI 2026-05 unverdicted novelty 6.0

Temporal reasoning is not the core bottleneck for LLMs on time-based QA; the real issue is unstructured text-to-event mapping, addressed by a neuro-symbolic system with PIS that reaches 100% accuracy on benchmarks whe...
Effective Performance Measurement: Challenges and Opportunities in KPI Extraction from Earnings Calls
cs.CL 2026-05 unverdicted novelty 6.0

Encoder models trained on SEC filings struggle with earnings calls due to domain shift, while LLMs enable open-ended KPI extraction with 79.7% human-verified precision on newly introduced benchmarks.
SBCA: Cross-Modal BERT-driven Actor-Critic for Multi-Asset Portfolio Optimization
q-fin.CP 2026-05 unverdicted novelty 6.0

SBCA is a reinforcement learning framework using BERT cross-modal fusion and Actor-Critic to integrate price data with sentiment text for multi-asset portfolio optimization with practical trading constraints.
SysTradeBench: An Iterative Build-Test-Patch Benchmark for Strategy-to-Code Trading Systems with Drift-Aware Diagnostics
cs.SE 2026-04 unverdicted novelty 6.0

SysTradeBench evaluates 17 LLMs on 12 trading strategies, finding over 91.7% code validity but rapid convergence in iterative fixes and a continued need for human oversight on critical strategies.
PolySwarm: A Multi-Agent Large Language Model Framework for Prediction Market Trading and Latency Arbitrage
cs.AI 2026-04 unverdicted novelty 6.0

PolySwarm aggregates predictions from 50 LLM personas for Polymarket trading using Bayesian combination and divergence metrics, outperforming single models in calibration while adding latency arbitrage via CEX price models.
AutoDAN: Generating Stealthy Jailbreak Prompts on Aligned Large Language Models
cs.CL 2023-10 conditional novelty 6.0

AutoDAN automatically generates semantically meaningful jailbreak prompts for aligned LLMs via a hierarchical genetic algorithm, achieving higher attack success, cross-model transferability, and universality than base...
BloombergGPT: A Large Language Model for Finance
cs.LG 2023-03 conditional novelty 6.0

BloombergGPT is a 50B parameter LLM trained on a 708B token mixed financial and general dataset that outperforms prior models on financial benchmarks while preserving general LLM performance.
Learning to Trade Like an Expert: Cognitive Fine-Tuning for Stable Financial Reasoning in Language Models
cs.LG 2026-04 unverdicted novelty 5.0

A new fine-tuning framework with textbook-derived MCQs and simulation-based testing enables smaller open LLMs to show competitive, risk-aware financial trading behavior that outperforms baselines.
Empirical Evaluation of PDF Parsing and Chunking for Financial Question Answering with RAG
cs.CL 2026-04 unverdicted novelty 5.0

Systematic tests show that specific PDF parsers combined with overlapping chunking strategies better preserve structure and improve RAG answer correctness on financial QA benchmarks including the new TableQuest dataset.
Persistent and Conversational Multi-Method Explainability for Trustworthy Financial AI
cs.AI 2026-05 unverdicted novelty 4.0

An architecture stores XAI explanations persistently in searchable storage and uses RAG to synthesize multiple methods conversationally, cutting hallucination rates by 36% in a FinBERT financial sentiment demo.
The Acoustic Camouflage Phenomenon: Re-evaluating Speech Features for Financial Risk Prediction
cs.SD 2026-04 unverdicted novelty 3.0

Acoustic features degrade NLP performance in predicting stock volatility from earnings calls, attributed to 'Acoustic Camouflage' from media-trained vocal regulation.
A Review of Large Language Models for Stock Price Forecasting from a Hedge-Fund Perspective
q-fin.PR 2026-04 unverdicted novelty 3.0

This review synthesizes LLM uses in stock forecasting and catalogs key practical pitfalls from a hedge-fund viewpoint.
Developing an ESG-Oriented Large Language Model through ESG Practices
cs.CE 2026-03 unverdicted novelty 3.0

ESG-adapted versions of Qwen-3-4B using LoRA and IRM outperform the base model and Llama-3/Gemma-3 baselines on generative ESG question-answering tasks.

Reference graph

Works this paper leans on

35 extracted references · 35 canonical work pages · cited by 18 Pith papers · 13 internal anchors

[1]

Basant Agarwal and Namita Mittal. 2016. Machine Learning Approach for Sentiment Analysis. Springer International Publishing, Cham, 21–45. https: //doi.org/10.1007/978-3-319-25343-5_3

work page doi:10.1007/978-3-319-25343-5_3 2016
[2]

Fernando Sánchez-Rada, and Carlos A

Oscar Araque, Ignacio Corcuera-Platas, J. Fernando Sánchez-Rada, and Carlos A. Iglesias. 2017. Enhancing deep learning sentiment analysis with ensemble tech- niques in social applications. Expert Systems with Applications 77 (jul 2017), 236–246. https://doi.org/10.1016/j.eswa.2017.02.002

work page doi:10.1016/j.eswa.2017.02.002 2017
[3]

Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2018. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. (2018). https://doi.org/arXiv:1811.03600v2 arXiv:1810.04805

work page internal anchor Pith review Pith/arXiv arXiv 2018
[4]

Li Guo, Feng Shi, and Jun Tu. 2016. Textual analysis and machine leaning: Crack unstructured data in finance and accounting. The Journal of Finance and Data Science 2, 3 (sep 2016), 153–170. https://doi.org/10.1016/J.JFDS.2017.02.001

work page doi:10.1016/j.jfds.2017.02.001 2016
[5]

Jeremy Howard and Sebastian Ruder. 2018. Universal Language Model Fine- tuning for Text Classification. (jan 2018). arXiv:1801.06146 http://arxiv.org/abs/ 1801.06146

work page internal anchor Pith review Pith/arXiv arXiv 2018
[6]

Neel Kant, Raul Puri, Nikolai Yakovenko, and Bryan Catanzaro. 2018. Prac- tical Text Classification With Large Pre-Trained Language Models. (2018). arXiv:1812.01207 http://arxiv.org/abs/1812.01207

work page internal anchor Pith review Pith/arXiv arXiv 2018
[7]

Mathias Kraus and Stefan Feuerriegel. 2017. Decision support from financial disclosures with deep neural networks and transfer learning.Decision Support Sys- tems 104 (2017), 38–48. https://doi.org/10.1016/j.dss.2017.10.001 arXiv:1710.03954

work page internal anchor Pith review Pith/arXiv arXiv doi:10.1016/j.dss.2017.10.001 2017
[8]

Srikumar Krishnamoorthy. 2018. Sentiment analysis of financial news articles using performance indicators. Knowledge and Information Systems 56, 2 (aug 2018), 373–394. https://doi.org/10.1007/s10115-017-1134-1

work page doi:10.1007/s10115-017-1134-1 2018
[9]

Xiaodong Li, Haoran Xie, Li Chen, Jianping Wang, and Xiaotie Deng. 2014. News impact on stock price return via sentiment analysis. Knowledge-Based Systems 69 (oct 2014), 14–23. https://doi.org/10.1016/j.knosys.2014.04.022

work page doi:10.1016/j.knosys.2014.04.022 2014
[10]

Bing Liu. 2012. Sentiment Analysis and Opinion Mining. Synthesis Lectures on Human Language Technologies 5, 1 (may 2012), 1–167. https://doi.org/10.2200/ s00416ed1v01y201204hlt016

work page 2012
[11]

Tim Loughran and Bill Mcdonald. 2011. When Is a Liability Not a Liability? Textual Analysis, Dictionaries, and 10-Ks. Journal of Finance 66, 1 (feb 2011), 35–65. https://doi.org/10.1111/j.1540-6261.2010.01625.x

work page doi:10.1111/j.1540-6261.2010.01625.x 2011
[12]

Tim Loughran and Bill Mcdonald. 2016. Textual Analysis in Accounting and Finance: A Survey. Journal of Accounting Research 54, 4 (2016), 1187–1230. https://doi.org/10.1111/1475-679X.12123

work page doi:10.1111/1475-679x.12123 2016
[13]

Bernhard Lutz, Nicolas Pröllochs, and Dirk Neumann. 2018. Sentence-Level Sentiment Analysis of Financial News Using Distributed Text Representations and Multi-Instance Learning. Technical Report. arXiv:1901.00400 http://arxiv.org/ abs/1901.00400

work page internal anchor Pith review Pith/arXiv arXiv 2018
[14]

Macedo Maia, Andrï£¡ Freitas, and Siegfried Handschuh. 2018. FinSSLx: A Senti- ment Analysis Model for the Financial Domain Using Text Simplification. In2018 IEEE 12th International Conference on Semantic Computing (ICSC) . IEEE, 318–319. https://doi.org/10.1109/ICSC.2018.00065

work page doi:10.1109/icsc.2018.00065 2018
[15]

Macedo Maia, Siegfried Handschuh, André Freitas, Brian Davis, Ross Mcdermott, Manel Zarrouk, Alexandra Balahur, and Ross Mc-Dermott. 2018. Companion of the The Web Conference 2018 on The Web Conference 2018, {WWW} 2018, Lyon , France, April 23-27, 2018. ACM. https://doi.org/10.1145/3184558

work page doi:10.1145/3184558 2018
[16]

Burton G Malkiel. 2003. The Efficient Market Hypothesis and Its Critics. Jour- nal of Economic Perspectives 17, 1 (feb 2003), 59–82. https://doi.org/10.1257/ 9 089533003321164958

work page 2003
[17]

Pekka Malo, Ankur Sinha, Pekka Korhonen, Jyrki Wallenius, and Pyry Takala

work page
[18]

Good Debt or Bad Debt: Detecting Semantic Orientations in Economic Texts

Good debt or bad debt: Detecting semantic orientations in economic texts. Journal of the Association for Information Science and Technology 65, 4 (2014), 782–796. https://doi.org/10.1002/asi.23062 arXiv:arXiv:1307.5336v2

work page internal anchor Pith review Pith/arXiv arXiv doi:10.1002/asi.23062 2014
[19]

G. Marcus. 2018. Deep Learning: A Critical Appraisal. arXiv e-prints (Jan. 2018). arXiv:cs.AI/1801.00631

work page internal anchor Pith review Pith/arXiv arXiv 2018
[20]

Justin Martineau and Tim Finin. 2009. Delta TFIDF: An Improved Feature Space for Sentiment Analysis.. In ICWSM, Eytan Adar, Matthew Hurst, Tim Finin, Natalie S. Glance, Nicolas Nicolov, and Belle L. Tseng (Eds.). The AAAI Press. http://dblp.uni-trier.de/db/conf/icwsm/icwsm2009.html#MartineauF09

work page 2009
[21]

Bryan McCann, James Bradbury, Caiming Xiong, and Richard Socher. 2017. Learned in Translation: Contextualized Word Vectors. Nips (2017), 1–12. arXiv:1708.00107 http://arxiv.org/abs/1708.00107

work page internal anchor Pith review Pith/arXiv arXiv 2017
[22]

Stephen Merity, Nitish Shirish Keskar, and Richard Socher. 2017. Regular- izing and Optimizing LSTM Language Models. CoRR abs/1708.02182 (2017). arXiv:1708.02182 http://arxiv.org/abs/1708.02182

work page internal anchor Pith review Pith/arXiv arXiv 2017
[23]

Jeffrey Pennington, Richard Socher, and Christopher Manning. 2014. Glove: Global Vectors for Word Representation. InProceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) . Association for Computational Linguistics, Doha, Qatar, 1532–1543. https://doi.org/10.3115/v1/ D14-1162

work page doi:10.3115/v1/ 2014
[24]

Matthew E Peters, Mark Neumann, Mohit Iyyer, Matt Gardner, Christopher Clark, Kenton Lee, and Luke Zettlemoyer. 2018. Deep contextualized word representations. (2018). https://doi.org/10.18653/v1/N18-1202 arXiv:1802.05365

work page internal anchor Pith review Pith/arXiv arXiv doi:10.18653/v1/n18-1202 2018
[25]

Guangyuan Piao and John G Breslin. 2018. Financial Aspect and Sentiment Predictions with Deep Neural Networks. 1973–1977. https://doi.org/10.1145/ 3184558.3191829

work page arXiv 2018
[26]

Aliaksei Severyn and Alessandro Moschitti. 2015. Twitter Sentiment Analysis with Deep Convolutional Neural Networks. InProceedings of the 38th International ACM SIGIR Conference on Research and Development in Information Retrieval - SIGIR '15. ACM Press. https://doi.org/10.1145/2766462.2767830

work page doi:10.1145/2766462.2767830 2015
[27]

Sahar Sohangir, Dingding Wang, Anna Pomeranets, and Taghi M Khoshgoftaar

work page
[28]

Journal of Big Data 5, 1 (2018)

Big Data: Deep Learning for financial sentiment analysis. Journal of Big Data 5, 1 (2018). https://doi.org/10.1186/s40537-017-0111-6

work page doi:10.1186/s40537-017-0111-6 2018
[29]

Chi Sun, Xipeng Qiu, Yige Xu, and Xuanjing Huang. 2019. How to Fine-Tune BERT for Text Classification? (2019). arXiv:1905.05583 https://arxiv.org/pdf/ 1905.05583v1.pdfhttp://arxiv.org/abs/1905.05583

work page arXiv 2019
[30]

Abinash Tripathy, Ankit Agrawal, and Santanu Kumar Rath. 2016. Classification of sentiment reviews using n-gram machine learning approach. Expert Systems with Applications 57 (sep 2016), 117–126. https://doi.org/10.1016/j.eswa.2016.03. 028

work page doi:10.1016/j.eswa.2016.03 2016
[31]

Attention Is All You Need

Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. Attention Is All You Need. Nips (2017). arXiv:1706.03762 http://arxiv.org/abs/1706.03762

work page internal anchor Pith review Pith/arXiv arXiv 2017
[32]

Casey Whitelaw, Navendu Garg, and Shlomo Argamon. 2005. Using appraisal groups for sentiment analysis. In Proceedings of the 14th ACM international conference on Information and knowledge management - CIKM '05. ACM Press. https://doi.org/10.1145/1099554.1099714

work page doi:10.1145/1099554.1099714 2005
[33]

Steve Yang, Jason Rosenfeld, and Jacques Makutonin. 2018. Financial Aspect- Based Sentiment Analysis using Deep Representations. (2018). arXiv:1808.07931 https://arxiv.org/pdf/1808.07931v1.pdfhttp://arxiv.org/abs/1808.07931

work page internal anchor Pith review Pith/arXiv arXiv 2018
[34]

Lei Zhang, Shuai Wang, and Bing Liu. 2018. Deep learning for sentiment analysis: A survey. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery 8, 4 (mar 2018), e1253. https://doi.org/10.1002/widm.1253

work page doi:10.1002/widm.1253 2018
[35]

Yukun Zhu, Ryan Kiros, Richard Zemel, Ruslan Salakhutdinov, Raquel Urtasun, Antonio Torralba, and Sanja Fidler. 2015. Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books. (jun 2015). arXiv:1506.06724 http://arxiv.org/abs/1506.06724 10

work page internal anchor Pith review Pith/arXiv arXiv 2015