Arkin Gupta

Don't miss out on an exciting case study at the 6th Annual Machine Learning in Quantitative Finance conference! Igor Halperin from Fidelity Investments will explore the applications of reinforcement learning techniques to optimize quantitative trading. Key topics include: 1. Understanding how reinforcement learning empowers algorithmic trading 2. strategies to adapt and evolve in real time. 3. Identifying best practices for applying reinforcement learning techniques to portfolio optimization. 4. Delving into recent developments in reinforcement learning and their impact on quantitative finance. Join us from September 23-25, 2024, in New York, NY, to examine the best use cases of machine learning techniques in quantitative finance strategies and discover actionable solutions to overcome implementation challenges. #MachineLearning #QuantitativeFinance #ReinforcementLearning #AlgorithmicTrading #FinanceConference #NewYork Register now to secure your spot! Event Website: https://lnkd.in/etABFXzP

19

Interesting strategy assuming the market is endogenous. Using deep #RL to dynamically adjust portfolio allocation to max. returns while managing risk. An #Autoencoder to extract 3 latent features from the initial set of 12. Continuously learning from new data based on the input distribution. Would be interesting to see the performance when the strategy is exposed to exogenous factors. Missing a public repo for this but still a great read! https://bit.ly/4bwV1zt

1

So right now Im studying derivatives, I read about credit default swaps and my mischiefs mind had a thought. Lets say we have a buyer of the CDS and there is a default event, and the buyer has it in their contracts with the defaulters that upon default the assets will be sold to someone else and the buyer reserves the right to buy it back at the same valuation as a way to reduce the paper value of the recovered debt. Would this be considered illegal or unethical, or is the buyer obliged to disclose this before hand? #financialshenanigans #CFA #derivatives

7

A behind-the-scenes look at how I structure Financial Models: ~~~ tl;dr: if you want the confidence that your model is structurally accurate, check this out: https://lnkd.in/eAmU_sgK ~~~ 1. Raw Data 2. Support Schedules 3. Three Statement Model 4. Summaries 5. Control Panel And yes, it's often very helpful to sketch things out like this: It saves you from building a bunch of wasted tabs you don't need. ~~~ ♻️Please reshare with your network if you found this helpful, thanks!

2

I am planning to post couple of interesting and award winning papers from #NeurIPS for next coupe of days:  Title: 𝗡𝗼𝘁 𝗔𝗹𝗹 𝗧𝗼𝗸𝗲𝗻𝘀 𝗔𝗿𝗲 𝗪𝗵𝗮𝘁 𝗬𝗼𝘂 𝗡𝗲𝗲𝗱 Selective Language Modeling for Efficient LLM Training Paper Link: https://lnkd.in/eXFBGDAU AI Generated Podcast: https://lnkd.in/eXH7CuX8 In the evolving landscape of AI research, this paper on Selective Language Modeling (SLM) offers intriguing perspectives on training efficiency. The RHO-1 approach introduces a nuanced method of language model training. By using a reference model to score token importance, the research challenges traditional data-intensive training paradigms. 𝗞𝗲𝘆 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗢𝗯𝘀𝗲𝗿𝘃𝗮𝘁𝗶𝗼𝗻𝘀: • Intelligent Token Selection: Researchers explored a method of focusing model training on tokens with higher informational utility • Data Efficiency Approach: Demonstrated potential for improved performance using fewer training tokens • Precision in Machine Learning: Highlighting the value of targeted training strategies 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁𝗮𝗹 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝘀: • Improvements in few-shot learning capabilities • Enhanced accuracy across mathematical and general task domains • Potential for more resource-efficient model development #MachineLearning #AIResearch #TechInnovation #NeurIPS

15

Do Artificial intelligence and machine learning work in the world of Quant investing?🤔 Join Devina Mehra as she breaks it down for us. She'll explain why asking this question might be missing the point, why not all quant models are the same, and what it really takes to make these models successful.🚀 Learn about the unique approach of First Global’s Human+Machine Model and how it sets a new standard in investing. If you’re curious about the real deal behind AI, ML, and quant investing, this video is a must-watch!📊🤖 #InsightsByDevina No. #31 Let’s dive in and get smarter about investing together! 📊💡

22

3 Comments

You may have heard RNN/LSTM and the time-series LLM Lag-Llama. You may be curious about the development of time series modeling from RNN/LSTM to Temporal Fusion Transformers and then Lag-Llama. Besides showing you how to apply the techniques, this book explain the evolution. I'd like to make this post, which was behind the medium.com paywall, open to you. It is Chapter 19 in the book "Modern Time Series Forecasting Techniques." Have you pick up a copy? The ebook is made available on a secure platform teachable.com for readers who prefer the digital edition. Sample eBook chapters: https://lnkd.in/e9JX-qKa eBook on Teachable: https://lnkd.in/eRsFBsYT Amazon print book: https://a.co/d/25FVsMx https://lnkd.in/eSEgnhF9

11

𝗖𝗮𝗻 𝗔𝗜 𝗧𝗲𝗹𝗹 𝗦𝘁𝗼𝗿𝗶𝗲𝘀 𝗟𝗶𝗸𝗲 𝗛𝘂𝗺𝗮𝗻𝘀? 𝗘𝗻𝘁𝗲𝗿 𝗠𝗶𝘅𝘁𝘂𝗿𝗲-𝗼𝗳-𝗣𝗲𝗿𝘀𝗼𝗻𝗮𝘀 (𝗠𝗼𝗣) A new strategy called 𝗠𝗶𝘅𝘁𝘂𝗿𝗲-𝗼𝗳-𝗣𝗲𝗿𝘀𝗼𝗻𝗮𝘀 (MoP) is redefining how Large Language Models (LLMs) evaluate stories. By prompting LLMs to adopt personas focused on specific storytelling metrics—like empathy, authenticity, or narrative complexity—MoP improves alignment with human judgment by over 20%. Psychological Depth Scale (PDS): A New Lens for Stories Using the PDS framework, researchers assessed the psychological depth of LLM-generated short stories based on: • Authenticity: Genuine human-like emotions. • Empathy: Evoking connection with characters. • Engagement: Captivating readers. • Emotion Provocation: Stirring strong reactions. • Narrative Complexity: Intricate plots and characters. 𝗞𝗲𝘆 𝗙𝗶𝗻𝗱𝗶𝗻𝗴𝘀 • LLMs rival humans: Models like GPT-4 created stories with depth comparable to human-written ones. • Size isn’t everything: Smaller models (e.g., Llama-2-7B) performed surprisingly well. • Emotional AI: MoP showed that LLMs could reflect diverse, nuanced human perspectives. 𝗪𝗵𝘆 𝗜𝘁 𝗠𝗮𝘁𝘁𝗲𝗿𝘀 This research shifts the focus from technical quality to emotional and intellectual engagement, opening doors for AI in education, entertainment, and even therapy. Can storytelling AI become the next creative frontier? What do you think? Paper Link: https://lnkd.in/ewq23aFA Spotify Episode of this paper: https://lnkd.in/enqEe5TZ #AI #Storytelling #GenerativeAI #PsychologicalDepth #MixtureOfPersonas #LLM #NotebookLM #Innovation

5

In context of debate if large context #LLM models will eventually negate the need for #RAG, an article from Databricks has some good insights - • Longer contexts boost RAG accuracy... to a point • Performance often peaks at 32k-64k tokens, then declines • Top models (GPT-4, Claude 3.5) maintain consistency at scale • Others show unique failure modes (e.g., copyright concerns, summarization instead of Q&A) • Optimal context size varies by model and task • Lack of long-context training data may explain some issues Key takeaway: Long context and RAG are synergistic, but LLMs need refinement to fully leverage extended contexts. https://lnkd.in/eehdKtCm

17

Before anyone read Kimball books or build a data warehouse, they should read Ralph Kimball fundamental article about ER Modeling vs Dimensional Modeling: https://lnkd.in/eD3ab7Ve (thanks to Keith Belanger for reminding all of us today). In this article, Ralph Kimball explained very clearly, why a normalised database such as 3NF should not be used for reporting and analytics. He listed all the issues if we use such database for data warehousing. I know there are exceptions, such as the cases that Golov Nikolai has showed us in his paper with Lars Rönnbäck (Anchor Modeling) but in the majority of cases, Ralph Kimball is right: Dimensional Modeling is far more superior than a normalised database for #datawarehousing, #reporting and #analytics. Ralph wrote in that article: "The resulting master DM model of a data warehouse for a large enterprise will consist of somewhere between 10 and 25 very similar-looking star join schemas. Each star join will have four to 12 dimension tables. If the design has been done correctly, many of these dimension tables will be shared from fact table to fact table." That is amazing. To be able to come up with those figures he must have done many implementations! 10-25 stars, 4-12 dims per star. And that was back in 1997! That's 27 years ago. Amazing. Every #DataArchitect needs to read this article. It is such a fundamental article that everyone in the #Data world needs to read. Thank you Ralph Kimball for this amazing gift: Dimensional Model.

185

35 Comments

Mathematical breakdown of different types of PCA with their implementation in Quant Finance https://lnkd.in/eZ2pkm3t 

22

When assessing the performance of Large Language models (LLMs) we see the primary focus being technical benchmarks, such as graduate-level reasoning, coding proficiency, and multilingual math capabilities. These metrics are undeniably important, but only paint part of the picture. The recent performance comparison of models like Claude 3.5 Sonnet, Claude 3 Opus, GPT-4o, and Gemini 1.5 Pro underscores this point. While Claude 3.5 Sonnet excelled in graduate-level reasoning, coding, and multilingual math, it begs the question: How long until we start incorporating soft skills and behavioural traits into our model assessments? Soft Skills: The Human Touch in AI [1] Problem-Solving: Beyond solving complex equations, AI should demonstrate strong analytical and problem-solving skills, much like what we expect from top-tier talent. This involves tackling intricate financial problems and developing innovative solutions. [2] Communication: The ability to communicate technical concepts clearly and effectively to both technical and non-technical stakeholders is crucial. AI models need to bridge the gap between data science and business operations, ensuring everyone is on the same page. [3] Adaptability: As technologies evolve, so should our models. Adaptability in learning new technologies, adjusting to changing environments, and handling multiple tasks simultaneously is a trait we value in human employees and should equally prioritise in AI. [4] Teamwork: Collaboration is key. AI models should facilitate and enhance teamwork, contributing to a positive and productive work environment. They should be designed to integrate seamlessly with human teams, complementing their efforts. [5] Attention to Detail: Accuracy and precision in financial data and software development are non-negotiable. AI must exhibit meticulousness in its operations, ensuring high standards of quality and reliability. Behavioural Traits: Ethical and Proactive AI [1] Motivation and Initiative: Just as we look for candidates who demonstrate a proactive attitude towards learning and self-improvement, our models should show a capacity for continuous improvement and an 'interest' in advancing the fintech landscape. [2] Ethical Judgement: Understanding ethical considerations, such as data privacy, fairness, and regulatory compliance, is paramount. AI models must be designed with these principles in mind, ensuring they operate within ethical boundaries. Integrating these soft skills and behavioural traits into our assessment criteria will provide a more holistic view of an AI model's capabilities. It's not just about excelling in isolated technical tasks but also about how well these models can adapt, communicate, and integrate into broader business contexts. #AI #SoftSkills

13

The #VIX is an important risk barometer that gets considerable attention in times of market stress, but probably more than it deserves. It is subject to convexity, skew and calendar effects, all of which were in play during a massive spike over the last week with risk markets in free fall. The ratio of values between the first two VIX futures contracts often proves a more stable guide for making historic comparisons. On this basis, this week’s price action registers as a garden-variety correction, not a 2018 #Volmageddon event or anything close to LTCM, 9/11, GFC, COVID, or any other major risk-off episode. Our State Street Systemic Risk index underscores this point – it barely moved. #STTInsights

1

The Poisson jump process: Critical to accurate pricing of options and other derivatives. Used by professional quants at every bank. Now you can too (with a few lines of Python): Tomorrow, 26,000+ subscribers will get a step-by-step code walkthrough that simulates asset prices using a jump process. It’s about 20 lines of Python and takes about 5 minutes to read. You can get it right in your inbox tomorrow, for free: https://lnkd.in/enF-wvzX

72

2 Comments

The SABR model (Stochastic Alpha, Beta, Rho) is a widely used model in financial engineering for modeling forward interest rates and volatility. It was developed by Patrick Hagan, Deep Kumar, Andrew Lesniewski, and Diana Woodward in their 2002 paper, "Managing Smile Risk." The model is particularly popular for its ability to handle different market conditions and effectively model the volatility smile in derivatives markets. Components of the SABR Model The SABR model includes four key parameters, each symbolized by a Greek letter: 1. α (Alpha): Represents the initial volatility of the forward rate. 2. β (Beta): Controls the elasticity of the volatility; essentially it determines the nature of the stochastic process underpinning the asset price. A beta of 1 corresponds to a geometric Brownian motion, while lower values allow for more general processes. 3. ρ (Rho): The correlation between the forward rate and its volatility. This is crucial as it dictates how much the two factors can influence each other. 4. ν (Nu): The volatility of volatility, or how erratic the changes in volatility are.

152

3 Comments

📰We were asked by #Reuters—in an article by Jaspreet Singh, "Applied Materials forecasts quarterly revenue above estimates on AI-driven demanded"—for thoughts on why Applied Materials (#AMAT) was down in post-market trading after reporting quarterly numbers today. ⚙OUR TAKE: We are probably seeing some aftermarket #profit taking in Applied Materials after a mostly positive #earnings report this afternoon. Corporate spending on #artificialintelligence is helping their growth—similar to what we’ve seen from other #tech players. 🖼BIG PICTURE: Overall there is not much to complain about in this report except for the rather wide $0.8 billion dispersion for the fourth quarter net sales forecast (of $6.53 billion to $7.33 billion) and the mixed bag of numbers from sales to China (down -24% quarter-over-quarter, but up +24% year-over-year). QUOTED EXCERPT "I think some of the aftermarket pullback is based on how much the stock went up today... the mixed sales report from China and the wide variance in their fourth-quarter forecast," Michael Ashley Schulman, chief investment officer at Running Point Capital, said. "Some analysts may have been looking for a more robust next-quarter forecast." ~~ “We are your #familyoffice” #investments #stocks #privateinvestments #wealthmanagement #AI https://lnkd.in/grdpusMw ~~ Disclosure: The opinions expressed are those of Running Point Capital Advisors, LLC (Running Point) and are subject to change without notice. This is not investment advice and this is not a recommendation to buy or sell a security. The opinions referenced are as of the date of publication, may be modified due to changes in the market or economic conditions, and may not necessarily come to pass. Past performance is not indicative of future results. Forward-looking statements cannot be guaranteed. Running Point is an investment adviser registered with the U.S. Securities and Exchange Commission. Registration does not imply a certain level of skill or training. More information about Running Point’s investment advisory services and fees can be found in its Form ADV Part 2, which is available upon request. RP-24-120

17

3 Comments

Got a question, Can SECTOR evaluate summarization use cases? Yes, as SECTOR breaks down summary and looks at relevant part of context to compare This Engineering + unique statistical metrics makes it suitable for all use cases. All of this without having a secondary LLM for evaluation Here is the example to evaluate summary - https://lnkd.in/gU-9hqf7 See a sample output below #generativeai ##artificialintelligence

11

6 Comments

Anthropic's October model releases of Claude 3.5 Sonnet and Haiku have shown remarkable performance improvements in financial analysis and transcript processing. Compared to their previous 3 series models, these new releases demonstrated significant advancements in Aiera's specialized benchmarks. Our Financial Q&A benchmark, which evaluates a model's ability to comprehend and perform complex calculations on dense financial text, revealed substantial gains. These performance increases highlight the models' enhanced capacity to capture and generalize intricate patterns within financial data. Interestingly, while the new models excelled in financial analysis, we observed a slight regression in sentiment analysis for the Haiku models. This finding underscores the critical importance of careful model selection based on specific task requirements. You can view these results as well as others on our Hugging Face leaderboard: https://lnkd.in/gedYBf_p And learn more about Aiera here: https://www.aiera.com/

35

### 𝗠𝗲𝗺𝗼𝗥𝗔𝗚: 𝗔 𝗥𝗲𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻𝗮𝗿𝘆 𝗥𝗔𝗚 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 𝗳𝗼𝗿 𝗘𝗻𝗵𝗮𝗻𝗰𝗲𝗱 𝗘𝘃𝗶𝗱𝗲𝗻𝗰𝗲 𝗥𝗲𝘁𝗿𝗶𝗲𝘃𝗮𝗹! MemoRAG is transforming the way we approach retrieval-augmented generation (RAG) by incorporating a super-long memory model for global understanding across massive datasets. 🌐 In real-world applications, memory plays a critical role. For instance, in a knowledge-intensive question-answering task, using MemoRAG improved response accuracy by 27% compared to traditional RAG frameworks. Such gains highlight the value of global memory in providing nuanced, contextually rich answers. #### 🔥 𝗞𝗲𝘆 𝗙𝗲𝗮𝘁𝘂𝗿𝗲𝘀 🔥 1. 𝗚𝗹𝗼𝗯𝗮𝗹 𝗠𝗲𝗺𝗼𝗿𝘆: Processes up to 1 million tokens in a single context, offering unmatched depth and breadth across data. 2. 𝗢𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝗯𝗹𝗲 & 𝗙𝗹𝗲𝘅𝗶𝗯𝗹𝗲: Fine-tune and adapt MemoRAG to new tasks with only a few hours of training! 3. 𝗖𝗼𝗻𝘁𝗲𝘅𝘁𝘂𝗮𝗹 𝗖𝗹𝘂𝗲𝘀: Bridges raw inputs to answers using clues derived from memory, unlocking insights from even the most complex queries. 4. 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗖𝗮𝗰𝗵𝗶𝗻𝗴: Up to 30x faster context pre-filling through advanced caching, chunking, indexing, and encoding. 5. 𝗖𝗼𝗻𝘁𝗲𝘅𝘁 𝗥𝗲𝘂𝘀𝗲: Long contexts can be encoded once and reused, improving efficiency for tasks with repetitive data needs. 🔄 𝗟𝗶𝘁𝗲 𝗠𝗼𝗱𝗲 𝗼𝗳 𝗠𝗲𝗺𝗼𝗥𝗔𝗚 Experience MemoRAG’s powerful pipeline with just a few lines of code! Ideal for GPUs with 16GiB or 24GiB memory, the Lite Mode simplifies getting started while maintaining exceptional performance. #### ✨ 𝗕𝗮𝘀𝗶𝗰 𝗨𝘀𝗮𝗴𝗲 & 𝗠𝗼𝗱𝗲𝗹 𝗖𝗼𝗺𝗽𝗮𝘁𝗶𝗯𝗶𝗹𝗶𝘁𝘆 MemoRAG works seamlessly with HuggingFace models, using the `MemoRAG.memorize()` method to build global memory across long input contexts. #### 🧠 Long LLMs as Memory Models MemoRAG supports long-context LLMs like Meta-Llama-3.1-8B-Instruct and Llama3.1-8B-Chinese-Chat, optimizing memory through MInference. #MemoRAG #RAGFramework #LLMs #GlobalMemory #AI #NLP #MInference #GenerativeAI #MasteringLLM

88

2 Comments