Anna Graber-Naidich, PhD, MSc

I’m excited to share that our work, "FaMTEB: Massive Text Embedding Benchmark in Persian Language", has been accepted to EMNLP 2025 (Findings)! 🚀 The lack of a comprehensive benchmark for evaluating text embeddings in Persian has been a major challenge. Our team at MCINEXT, with the goal of addressing this gap, introduced FaMTEB, a benchmark for evaluating text embeddings in Persian, which is now featured on the reputable MTEB leaderboard. I’m grateful to have collaborated with Erfan Zeinivand and Mehran Sarmadi on this project. A special thanks to everyone who shared their knowledge and offered support throughout this journey. 🔗 Explore the leaderboard: https://lnkd.in/dHfmhzW7 📄 Read our paper: https://lnkd.in/dwWVbsD4 #EMNLP #EMNLP2025 #MCINEXT #FaMTEB #MTEB #PersianNLP

16

🔎 Making Data Simple with Principal Component Analysis (PCA) In today’s world, we often face large datasets with dozens of variables. How do we make sense of all this information without losing meaning? Use Principal Component Analysis (PCA). 🚀 Why PCA is Powerful ✨ Reduces complexity: turns many variables into fewer “principal components.” ✨ Preserves key information: keeps most of the variance in the data. ✨ Reveals hidden structure: uncovers patterns that aren’t obvious. ✨ Improves efficiency – simplifies models, reduces noise, and speeds up analysis. 👉 Bottom line: PCA helps us see the “big picture” by condensing complex data into clear, actionable insights. #DataScience #Statistics #MachineLearning #PCA #BigData

1

As #WWDC 2025 kicks off today, Apple’s recent paper - The Illusion of Thinking - is polarizing the AI community. Parshin Shojaee, Iman Mirzadeh, keivan alizadeh vahid, Maxwell Horton, Samy Bengio & Mehrdad Farajtabar's paper shows that reasoning models (LRMs) lag behind LLMs on simple tasks, excel at medium complexity, but fail completely on highly complex problems. LLMs/LRMs are INCREDIBLY useful for some tasks. Just revolutionary. But they are not a straight line to AGI & we need new AI architectures beyond just scaling up models as Yann LeCun, Gary Marcus & other AI leading voices have pointed out for a while now. Full take in today’s AI World stories 👇

15

On the Duolingo English Test, we're constantly researching better methods for automatically assessing writing and speaking. Check out our latest paper on automatic pronunciation scoring!

9

🎥 New ML Tutorial in Persian — Session 7 is live! In this video, I walk through how bagging + TruncatedSVD can help overcome overfitting in real-world datasets. We use shallow decision trees and dimensionality reduction to boost generalization and reduce variance. 📌 Dataset: Adult Income 📌 Tools: BaggingClassifier, TruncatedSVD, scikit-learn 📌 Language: Persian 📌 Audience: Teens, beginners, and curious learners Watch here 👉 https://lnkd.in/dh4YBTJv #MachineLearning #Bagging #TruncatedSVD #Overfitting #MLForTeens #PersianML #ScikitLearn #DimensionalityReduction #Generalization #AIeducation #MLTutorial #AdultIncomeDataset

8

4 Comments

I’m excited to share that our scoping review preprint has been published on medRxiv: A Scoping Review of Algorithmic Equity, Data Diversity, and Inclusive Design in the Transformer Era of Clinical NLP https://lnkd.in/eWvWM8h6 Many thanks to Abeer Badawi, Ph.D., Elham Dolatabadi, and Farah Ahmad, MBBS, MPH, PhD, for their collaboration, guidance, and support throughout this work. The rapid digitization of healthcare has positioned transformer-based NLP models as powerful tools for managing clinical text. However, their growing integration into practice raises important and unresolved questions about equity and inclusivity. This review synthesizes 56 studies (2017–2024) and examines how equity is addressed across three dimensions: ➡️ Algorithmic equity — fairness audits are largely post hoc and fragmented ➡️ Data diversity & representativeness — persistent underrepresentation creates what we define as Data Diversity Debt ➡️ Participatory design — observed in only 11% of studies, revealing a major gap beyond clinician involvement. To move beyond descriptive audits, we propose an equity-by-design roadmap to embed fairness, inclusivity, and accountability across the full lifecycle of clinical NLP systems—and to retire Data Diversity Debt. #HealthAI #ClinicalNLP #AlgorithmicEquity #DigitalHealthEquity #AIethics #ResponsibleAI #medRxiv

26

2 Comments

🧪 Sam Altman and Lewis André just launched a new company — positioned as “a new kind of R&D organization.” The company, Episteme, plans to hire exceptional researchers across physics, biology, computing, and energy — and give them something extremely rare: freedom to pursue long-horizon breakthrough ideas without bureaucracy or commercial pressure. What Episteme promises: 🧠 World-class researchers from diverse fields 🏗️ Full resources, infrastructure, and support 💡 Freedom to explore bold, high-impact scientific directions 📈 Equity in the company for every researcher 🚫 No short-term product deadlines or corporate KPIs The motivation is clear: Academia underfunds risky long-term research, and industry deprioritizes it because it doesn’t produce quarterly returns. Episteme wants to fill that gap — focusing on areas that could reshape both science and the global economy: physics, biology, computation, energy. The announcement sounds almost too good to be true — a kind of modern Bell Labs, but privately run and equity-driven. The big question now: who will join them? Website: episteme.com #Science #Research #Innovation #DeepTech #SamAltman #R&D #FutureOfScience #Physics #Biology #Energy #Compute

2

#ODSC Open Data Science Conference (ODSC) Today, I attended an insightful 6-hour workshop on building AI agents using the OpenAI Agents SDK and CrewAI. The hands-on session covered how to define, design, and develop agentic systems in Python, exploring frameworks like OpenAI Agents SDK, CrewAI, and MCP (Model Context Protocol). We built autonomous agents capable of tasks like code generation and autonomous trading. Key takeaways: - OpenAI's Agent SDK offers a lightweight interface for building and tracing agents. - CrewAI enables structured design using multi-agent teams. - MCP provides a standard for integrating tools and agents at scale. Thanks to the presenters, Ed Donner and Jon Krohn, for such a practical, forward-looking workshop. GitHub repo:

15

1 Comment