Jason Geng

This paper "ADVANCES AND CHALLENGES IN FOUNDATION AGENTS FROM BRAIN-INSPIRED INTELLIGENCE TO EVOLUTIONARY, COLLABORATIVE, AND SAFE SYSTEMS" surveys the current landscape of intelligent agents built on large language models (LLMs). It outlines a brain-inspired modular framework for agents, covering key functions like memory, perception, and planning. It explores how agents can evolve through self-improvement (e.g., AutoML), collaborate in multi-agent systems, and ensure safety and ethical alignment. By integrating insights from cognitive science, neuroscience, and computational research, the paper identifies key research gaps and challenges, encouraging innovations that balance technological advancement with societal benefits.​ Link: https://lnkd.in/d5g8C5DH Github: https://lnkd.in/dChZm_nw

12

📣 Today we are releasing ORQ, a new system we built from the ground up to enable complex analytics on large private datasets using cryptographically secure multiparty computation (MPC). For the first time ever, ORQ completes the notoriously challenging TPC-H benchmark, entirely under MPC, at a scale that had previously been achieved only with information leakage or the use of hardware enclaves. * Paper: https://lnkd.in/dvynBt-H * Code: https://lnkd.in/df2egqqm ORQ features: * Strong security: ORQ does not leak any information about the data to computing servers. It is designed to run on untrusted infrastructure, supports end-to-end oblivious execution, and protects all input, output, and intermediate data with provable security guarantees. * High expressivity: ORQ can securely compute a broad class of relational analytics, including database queries with multiple joins and complex aggregations that have been considered hard for cryptography. Users express analytics as dataflows, which are compiled into secure MPC programs by the ORQ engine. * Generality: ORQ makes black-box use of cryptographic primitives; therefore, it can be instantiated with any MPC protocol to support use cases with different threat models and security requirements. To showcase generality, we have instantiated ORQ with three state-of-the-art protocols: ABY (NDSS'15), the protocol by Araki et al. (CCS'16), and Fantastic Four (Security'21). * Scalability: ORQ employs a new approach to secure analytics that evaluates all queries we found in related works with asymptotic cost equal to that of sorting the input tables. It computes all TPC-H queries at Scale Factor 10 in a few hours, and other queries we collected from real MPC use cases in less than 10 minutes, on millions of input records. Real-world impact: ORQ is a big milestone for CASPLab and the result of a broader effort led by the Systems, Security, and PL groups at BU to push cryptographic computing beyond niche applications. We recently deployed ORQ in a pilot use case to create the first Workforce Data Report in Massachusetts, published in May 2025. Read more about this initiative here: https://lnkd.in/dbiZjTpB Kudos to my students Eli Baum, Sam Buxbaum, and Nitin Mathai for leading this work, and to Muhammad Faisal for developing the vectorized execution engine ORQ is based on. We will present the paper next month at SOSP'25. See you in Seoul! #SOSP #MPC #analytics #cryptography

119

2 Comments

How do large language models interpret words relating to probability like “unlikely,” “probably,” or “almost certain"? The below shows what happens when we compare judgements from different models to a benchmark dataset of human judgments (data from: https://lnkd.in/enByUDTx).

117

4 Comments

New short course: DSPy: Build and Optimize Agentic Apps DSPy is a powerful open-source framework for automatically tuning prompts for GenAI applications. In this course, you'll learn to use DSPy, together with MLflow. This is built in partnership with Databricks and taught by Chen Qian, co-lead of the DSPy framework. Many AI builders spend hours hand-tuning prompts. When given a set of evals, DSPy automates this process. It’s especially useful for optimizing prompts, including few-shot prompts, in complex agentic AI workflows. Further, if you switch an application to a newer LLM, performance can degrade if your prompts were optimized to the previous model. DSPy automatically optimizes the entire system for the new LLM as well, using just a few evaluation examples. This course teaches DSPy works, and best practices for using it. You’ll write programs using DSPy’s signature-based programming model, debug them with MLflow tracing -- to gain visibility into how different parts of a pipeline, as well as how the overall system, are performing -- and automatically improve their accuracy with DSPy Optimizer. Please sign up here: https://lnkd.in/gdjae8AX

4,671

190 Comments

Tuning the step size for gradient-based samplers is still one of the biggest practical headaches in Bayesian computation. For Metropolis-Hastings-based algorithms like the random-walk, MALA, HMC, etc., you can tune the step size based on an optimal acceptance rate. But, in machine learning, it's more common to use non–Metropolis–Hastings algorithms like ULA, SGLD, SVGD, etc. Here, a “good” step size is crucial: Too small → painfully slow mixing and wasted compute Too large → numerical instability and bias And in high dimensions, the usual rules of thumb break down quickly In a new paper led by Louis Sharrock, we tackle this problem by viewing sampling as an optimisation problem on the space of probability measures and designing a tuning-free step size schedule directly from a functional upper bound. E.g., a non-asymptotic bound on the Kullback-Leibler divergence. The result is FUSE: an adaptive step size rule that you can plug into ULA, SGLD, SVGD and other related Wasserstein gradient-flow methods to eliminate manual step size tuning, while retaining (almost) the same guarantees as if you had chosen the optimal constant step size in hindsight. We've written a blog post that explains the key ideas, with figures and intuition: 👉 Blog: https://lnkd.in/ee9K9biX And the full technical details are in the paper: 📄 Sharrock & Nemeth (2025), Tuning-Free Sampling via Optimization on the Space of Probability Measures https://lnkd.in/enwch4NE If you work with Langevin-based samplers, SGLD for large-scale Bayesian inference, or particle methods like SVGD and are tired of endlessly fiddling with learning rates, you might find FUSE a useful perspective — and, hopefully, a practical solution. #BayesianInference #MonteCarlo #MCMC #ProbabilisticAI #Optimization #MachineLearningResearch

105

2 Comments

🚀 ML Project to Practice 📊 Student Performance Prediction Where to Practice (Run the Project): 📚Google Colab Notebook: https://lnkd.in/gzaViEh4 ❗ Problem: There are many reasons students might perform poorly in their studies, and perhaps we can prevent it. 🤔 🎯 Goal: Predict if a student might perform badly and help them achieve better scores. 📈 🛠️ Approach: Use Linear Regression and evaluate how good our model is using Loss Functions. 📊 -------------------------- 🚀 Master Machine Learning 🔬 A Physics-Based Approach 💼 Get Hired or Promoted 📈 👉 Here’s your roadmap: 1. 🔗 Visit the program page: https://lnkd.in/gWwq78Vp 2. 🎬 Start from Episode 1 and move through the episodes step by step 👣. 1. 🔹 Read the Research Notes and watch the embedded videos to deeply understand each concept. 2. 🔹 Take the Course. Open the corresponding Colab notebook, run the code from top to bottom, and see how these physics-based ideas map onto conventional ML algorithms. 3. 🔹 Run the Project. Apply your knowledge on real datasets, solve real-world problems, and gain hands-on experience that bridges theory and practice. 4. 🌟 Ready to go even further? ✨ Turn the project into a unique-to-you professional GitHub repository — with a clean, industry-standard folder structure, production-ready settings, and depoly as a web application to showcase your work live. 📩 Click Request Professional Setup on the page to get started. 🚀 FREE-or-Land-Job 6 Months BootCamp From Graduation in a Math-Based Field to a Data Science Career ✨ Program Highlights 💼 Land a Job or Pay Almost Nothing! 🧭 Step-by-Step Roadmap: 100% done-for-you structured schedule outlining exactly what to learn at each stage. 🧑‍🏫 Weekly Live Sessions: Expert-led project-solving sessions to master core data science skills through hands-on experience. 🌐 Build Your Online Presence & Portfolio: Work on real-world projects, receive expert reviews, and showcase your work on GitHub. 🧠 Technical & Soft Skills: Master all skills needed to land a job, from machine learning to interview and resume prep. 📄 Job Application Prep & Mentoring: Step-by-step guide on job search strategies, networking, and application materials. 🤝 Get Support When Needed: Connect with mentors for help, ask questions anytime, and join group discussions to learn and grow together. 🎯 Apply now: 🔗 https://lnkd.in/gDv64aFg

5

In collaborative machine learning (CML), early contributors take on more risk and encourage participation from wait-and-see parties, but existing CML frameworks treat every party as if they join simultaneously. The #NeurIPS2025 paper of Jiangwei Chen, Nguyen Pham, Rachael Sim, Arun Verma, Zhaoxuan Wu, Chuan Sheng Foo, and Bryan Kian Hsiang Low proposes a time-aware CML framework that rewards parties not just for what they contribute, but when they contribute: •⁠ ⁠8 incentives extending fairness with time-aware notions that incentivize early participation while preventing low-quality early submissions. •⁠ ⁠2 reward distribution methods that theoretically satisfy all incentives and recover the #ShapleyValue in time-agnostic cases. •⁠ ⁠A practical recipe for constructing valuation functions and reward realization, with empirical validation on synthetic and real-world data. Find our more at our posters! #NeurIPS2025 Thu, Dec 4, Exhibit Hall C, D, E #1104 #EurIPS2025 Wed, Dec 3, Poster Stand #56 Paper: https://lnkd.in/g_yM6j39 #FederatedLearning

50