Efi Pylarinou’s Post

A very insightful, detailed and efficiency inducing blog post. I can see this used in enquiries, root cause analysis where detailed data requires summarising for exec meetings, elevator pitches and business unit updates. Nice work, Dinesh kumar Subramani and Pranav Sharma. Summary; Organizations across industries are using automatic text summarization to more efficiently handle vast amounts of information and make better decisions. In the financial sector, investment banks condense earnings reports down to key takeaways to rapidly analyze quarterly performance. Media companies use summarization to monitor news and social media so journalists can quickly write stories on developing issues. Government agencies summarize lengthy policy documents and reports to help policymakers strategize and prioritize goals. By creating condensed versions of long, complex documents, summarization technology enables users to focus on the most salient content. This leads to better comprehension and retention of critical information. The time savings allow stakeholders to review more material in less time, gaining a broader perspective. With enhanced understanding and more synthesized insights, organizations can make better informed strategic decisions, accelerate research, improve productivity, and increase their impact. The transformative power of advanced summarization capabilities will only continue growing as more industries adopt artificial intelligence (AI) to harness overflowing information streams. #ml #machinelearning #dataanalysis #bertscore #meteor #rouge

My latest article in Towards Data Science on productionizing graph RAG---like every worthwhile thing in data science and ML, it's not flip-a-switch easy! The article won't solve all of your graph RAG problems, but it might throw some blame at some worthy underlying causes...

As AI evolves, so does its ability to tackle complex scenarios. Graph RAG systems are at the forefront of this transformation, promising to enhance how we interconnect and utilize knowledge across industries. Read more from Brian Godsey, PhD now! #RAG

Why You Shouldn't Blindly Trust RAG Techniques. Under the name "Advanced RAG," various techniques are emerging, and many articles compare these techniques based on benchmarks. However, blindly believing and applying them is not recommended. Currently, there are no best practices, and the criteria for what is considered good can depends on the type and form of datasets and the nature of the RAG responses. During our team’s process of improving SuperSearchX, which answers finance-related questions, we conducted several experiments and found some interesting points (Of course, our test set itself may be biased, and our experiments were tailored to our specific use case.) 1. The embedding models listed on the Hugging Face leaderboard showed no significant differences. Surprisingly, even the open-source embedding models specialized for the finance domain had lower retrieval performance (Recall, F1) in our use case. 2. Techniques like HYDE and Sentence-Window Retrieval also did not yield significant results. 3. Experimenting with Data Chunking Strategies led to more meaningful performance improvements than other complex techniques. The conclusion here is that it’s difficult to find the right answer without direct experimentation and evaluation. To effectively run RAG, it's crucial to invest in the foundational infrastructure (Auto Evaluation Environment, LLMOps, etc.) early on so that various experiments can be conducted according to the use case. We have many more insights to share, and if you’re interested, why not join us directly? At Project Pluto, we are leading the way in the Finance AI Tech field and are looking for software engineers to join our journey. Interested engineers can check the job posting in the comments below!

AI poised to replace entry-level positions at large financial institutionshttps://lnkd.in/gadaf8_8 banks and investment firms are currently mulling plans to replace entry-level financial analyst positions with artificial intelligence (AI), with as many as two-thirds of these positions potentially on the chopping block. This anticipated move could completely transform how these companies hire new employees and how they manage and deliver the technology employees use. Major financial services organizations such as Goldman Sachs and Morgan Stanley are testing AI tools that could one day largely replace junior analysts and the hours of time they put into compiling and interpreting reports and crunching numbers and statistics — work that generative AI tools can do in a matter of seconds, according to a New York Times report . Large banking firms are quietly testing AI tools under code names such as as Socrates that could one day make the need to hire thousands of college graduates at these firms obsolete, according to the report. Indeed, Goldman Sachs has estimated that around 300 million workers could be significantly impacted by AI , and financial services companies appear to be moving in the direction to make this prediction a reality. Financial services organizations already use many kinds of automation technologies, from venerable workload automation tools and basic job schedulers each used by around half of all organizations, to more modern ones such as intelligent document processing or intelligent process automation using AI, used by 27% and 33% respectively, according to SMA Technologies’ report, The State of Automation in Financial Services 2024 . One of the key reasons why AI may disrupt entry-level financial services positions is its potential to automate the information processing that comprises a large portion their job description. Moreover, many of these employees experience burnout due to the “long hours and grunt work” that these positions entail, making it a less attractive option for college graduates, despite the significant salaries offered, according to an earlier New York Times report . Representatives from Goldman Sachs, JP Morgan Chase, and Morgan Stanley did not immediately respond to requests for comment on their companies’ plans to implement AI or its potential to change their hiring strategies. Transforming the workplace There is no doubt that generative AI already is transforming the workplace , but it remains to be seen just how much it will impact the jobs that humans do. Right now most organizations tend to be in the experimental phases of using the technology to supplement employee tasks, but that is likely to change, and quickly, experts say. As the financial industry is poised to be among the first to adopt AI in a big way, CIOs and other employees in this industry also likely will be the first to see it impact

Project -08 🌐 Event-Driven Price Movement Predictor 📈 🌍 Project Overview: Global events—like interest rate changes, geopolitical tensions, or natural disasters—significantly impact stock markets. But how do we quantify this impact? My solution uses AI to analyze real-time news, understand the sentiment and event context, and predict BUY/SELL signals for stocks. ⚙️ How It Works: 1️⃣ News Data Extraction 📰: -Scrapes stock-related news headlines using Google News and extracts event features. -Expands search across multiple pages for comprehensive coverage. 2️⃣ Event Feature Analysis 💬: -Processes headlines using TextBlob for sentiment analysis, capturing nuances like magnitude and polarity of events. -Converts raw news into actionable insights. 3️⃣ Historical Data Mapping 🗓️: -Maps these insights to historical stock prices (via Yahoo Finance) to understand past market reactions to similar events. 4️⃣ Predictive Model 🤖: -Trains a TensorFlow neural network to predict stock price movement probabilities. -Input: Event sentiment and historical price data. -Output: BUY or SELL recommendations based on market simulation. 📊 Visualization & Insights: -Clear Plots: Combines stock prices, sentiment trends, and trading signals. Tabular Summaries: Displays each news headline, its sentiment score, and the corresponding signal (BUY/SELL) in a detailed table. -Comprehensive Metrics: Tracks prediction accuracy and market performance. 💻 Technologies Used: -Google News 📰 for real-time data extraction. -GPT-4 🤖 for contextual analysis and event classification. -TensorFlow 🔗 for predictive modeling and trading simulations. -Matplotlib 📊 for rich, insightful visualizations. 🎯 Why This Project? As someone passionate about AI, finance, and data science, this project allowed me to explore the intersection of these fields. By combining real-time event tracking with predictive analytics, I aim to create a system that helps traders navigate the volatile markets with confidence. 💡 Looking Ahead: This is just the beginning! Future iterations could include: 📡 Real-time API integration for faster predictions. 🌐 Broader event categories (e.g., climate, political shifts). 🤝 Collaboration opportunities with finance professionals. 👋 Let’s Connect! I’d love to hear your feedback, thoughts, or ideas for improvement. If this aligns with your interests or you’d like to collaborate, let’s chat! 😊 #AI #Finance #MachineLearning #QuantFinance #StockMarket #DeepLearning #Python #DataScience

✅💡 What is QUALITY INTELLIGENCE? Discover how this transformative approach optimizes testing effectiveness and ensures high-quality software delivery. 👨🏻💻 This new blog post by Cristiano Caetano will give you the answers. 🚀 Read it here: https://lnkd.in/eEf9EeVp #QualityIntelligence #SoftwareTesting #Innovation #Testing #AI #AItesting

Twenty years ago, Armando Gonzalez shared a vision with me: a world where AI empowers decision-making through data. It was bold, it was pioneering—and today, he’s made it a reality. RavenPack’s launch of Bigdata.com isn’t just a platform; it’s a game-changer for financial research and decision-making. 🚀 Bigdata.com brings together billions of financial documents into a real-time, conversational AI—an incredible tool that lets professionals access, analyze, and act on critical insights at lightning speed. Whether it's summarizing earnings reports, tracking portfolios, or rebalancing positions in response to market shifts, this platform has redefined the way we work with data in finance. In an industry where AI tools often fall short, Bigdata.com stands out. It’s built specifically for finance, backed by leading investors, and is already transforming the research process for top institutions worldwide. This isn’t just a milestone for Armando and the RavenPack team—it’s a leap forward for the entire industry. Congratulations, Armando Gonzalez, on this monumental achievement! Your vision and relentless dedication are setting new standards for AI in finance. #Innovation #FinanceAI #BigData #AI #RavenPack #Leadership #DataRevolution #FinancialTech #ProudMoment https://lnkd.in/er4JpZZR

Learn how to compare and unify distinct data sets, transforming raw data into actionable insights. Crawlbase's latest article covers manual, semi-automatic, and automatic methods. Explore the essentials of data matching for web scraping, tailored for both businesses and individuals. #DataScience #WebScraping #DataMatching #BigData #AI #Tech #DataIntegration 

📊 The Importance of Data-Driven Decisions for Financial Advisors    Hello, Team! In today’s dynamic financial landscape, making informed, data-driven decisions is crucial for financial advisors. Relying on data helps eliminate biases, provides deeper insights, and enables more accurate predictions. Here’s why data-driven decision-making is essential and how AI can enhance this process.    Why Data-Driven Decisions Matter:  Accuracy and Precision: Data provides a factual basis for decisions, reducing the risk of errors.    Unbiased Insights: Data-driven analysis helps eliminate personal biases and emotions from the decision-making process.    Trend Analysis: Analyzing historical and real-time data reveals patterns and trends that inform future strategies.    Risk Management: Data helps identify and mitigate risks by highlighting potential issues before they escalate.    Client Confidence: Data-backed decisions build client trust and confidence, as they see decisions are based on solid evidence.    How AI Enhances Data-Driven Decision Making:  1. Advanced Analytics: AI tools like IBM Watson Analytics and Bloomberg Terminal analyze vast datasets to uncover hidden patterns and trends, providing financial advisors with actionable insights.    2. Predictive Modeling: AI uses historical data to predict future market movements. Tools like Kavout and Alpaca offer predictive analytics that help advisors anticipate market changes and adjust strategies accordingly.    3. Sentiment Analysis: AI can analyze news, social media, and other sources to gauge market sentiment. Platforms like MarketPsych and Sentifi provide insights into how current events may impact investments.    4. Real-Time Data Processing: AI processes data in real-time, offering up-to-date information that’s critical for timely decision-making. This is particularly valuable in fast-moving markets where quick decisions can make a significant difference.    5. Enhanced Research: AI tools like AlphaSense and Quid automate research by collecting and analyzing financial documents, news, and reports. This ensures advisors have the most relevant and current information at their fingertips.    Conclusion:  Data-driven decisions are essential for financial advisors aiming to stay competitive and deliver superior service to their clients. AI tools enhance this process by providing advanced analytics, predictive modeling, sentiment analysis, and real-time data processing.    If you’re a financial advisor looking to enhance your decision-making with AI and data-driven insights, contact us at "Invest in Yourself by Learning to Leverage AI." We offer personalized learning sessions and consulting services to help you integrate AI into your decision-making processes. Reach out today to learn more!    #AI #FinancialAdvisors #DataDriven #DecisionMaking #PredictiveAnalytics #RiskManagement #ClientConfidence #InvestmentStrategies #MarketTrends #RealTimeData #FinTech #BusinessGrowth