Scientific Blog

Learn, build, apply: free, hands-on resources where AI meets urban data science, from vision-language models, local LLMs and diffusion pipelines to geospatial analysis, with code you can run, a lecture series on AI, and critical reviews of published papers.

  • The AI Reading Series · Lecture 3: How Language Models Learned to Use Tools

The AI Reading Series · Lecture 3: How Language Models Learned to Use Tools

October 6, 2026|Getting Started, theory|

A language model can't check the weather, multiply two numbers reliably, or book a table, it can only predict text. Lecture 3 of the AI Reading Series shows how that limitation was turned into a strength: describe a tool in text, let the model emit a call, let ordinary code execute it. We trace the idea from RAG and WebGPT and prepare you to read Schick et al.'s "Toolformer" with the right questions in hand.

  • The AI Reading Series · Lecture 2: How Connectionism Conquered Language

The AI Reading Series · Lecture 2: How Connectionism Conquered Language

August 11, 2026|Getting Started, theory|

Part 2 of the AI reading series. The previous post ended with AlexNet's 2012 earthquake in image recognition. This one tells the road to language: how researchers turned words into vectors, taught networks to read sequences, discovered attention — and why, in 2017, eight Google researchers decided attention was all you need.

  • UVLM v3.1.0 — Qwen3-VL Joins the Registry, With Family-Based Model Selection

UVLM v3.1.0 — Qwen3-VL Joins the Registry, With Family-Based Model Selection

August 10, 2026|Advanced, Package, Python, Vision Language Model|

UVLM v3.1.0 adds a third model family, Qwen3-VL (2B–32B Instruct), bringing the registry to 15 checkpoints. The notebooks gain a two-level family/model selector, the loader picks BF16 automatically on capable GPUs, and the smallest new model runs in about 2 GB of VRAM. Same three-block workflow, same prompts, one more family to compare.

  • Deploy Your Own Local LLM on Low VRAM in 30 Minutes — A Private Chat Assistant in Jupyter

Deploy Your Own Local LLM on Low VRAM in 30 Minutes — A Private Chat Assistant in Jupyter

June 2, 2026|Advanced, Python|

Run a capable large language model entirely on your own machine — private, offline, and with as little as 8 GB of GPU memory. This hands-on guide sets up a clean Python environment, gets CUDA working even on the newest NVIDIA Blackwell cards, loads a 4-bit quantized model from Hugging Face, and builds an interactive chat widget with conversation memory and a live VRAM gauge in JupyterLab. No cloud, no API keys, no data leaving your computer.

  • SAGAI v2.0 — A Unified Multi-Model Notebook for Streetscape Analysis

SAGAI v2.0 — A Unified Multi-Model Notebook for Streetscape Analysis

May 21, 2026|Advanced, Python, Vision Language Model|

SAGAI v2.0 consolidates the full streetscape analysis pipeline into a single Google Colab notebook and replaces the inline LLaVA-only inference code with the UVLM package, enabling multi-model benchmarking across 11 VLM checkpoints. New features include a multi-task prompt builder, consensus validation with majority voting, chain-of-thought reasoning, truncation detection, interactive Folium maps, view-direction filtering, and support for loading existing polygons as study area boundaries.

  • Introducing UVLM: A Free Tool to Compare AI Models That Understand Images

Introducing UVLM: A Free Tool to Compare AI Models That Understand Images

March 17, 2026|Intermediate, Python, Vision Language Model|

UVLM is a free, open-source tool for loading, testing, and comparing Vision-Language Models on custom image analysis tasks. Running entirely in Google Colab, it lets researchers and practitioners benchmark multiple AI models using the same prompts and images — no coding, no GPU ownership, no model-specific pipelines. This post explains what VLMs are, why comparing them matters, and how to get started in five minutes.

  • From Large Language Models to Autonomous AI Agents — Architecture, Capabilities, and Emerging Risks

From Large Language Models to Autonomous AI Agents — Architecture, Capabilities, and Emerging Risks

February 19, 2026|Advanced|

Large Language Models are stateless, single-pass prediction engines — powerful but passive. Wrapping them in a perception–action loop with environment access and tool use transforms them into something qualitatively different: autonomous AI agents. This post walks through the transformer architecture, explains how the agent paradigm introduces closed-loop reasoning over environments and tasks, surveys the growing toolkit ecosystem, and examines the emerging risk landscape.