Hugging Face Unveils Microduck: A $399 Open-Source 25 cm Biped You Train with Reinforcement LearningMichal Sutter writes about Pollen Robotics, a Bordeaux-based team at Hugging Face, which has opened pre-orders for Microduck, a 25 cm bipedal robot priced at $399. Unlike most robotics launches that rely on demo videos, Microduck ships with its full training loop — every movement (walking, sitting, kicking, roller-skating, self-recovery) is a neural policy trained in a physics simulator and exported to hardware. The robot carries 15 motors, a camera, LiDAR, two IMUs, and a Rockchip RK3566, with policies trained via PPO in MuJoCo Warp in roughly one to two hours on a CUDA GPU.
- Sim-to-real hinges on a BAM actuator model (voltage control law, back-EMF, Coulomb/Stribeck/load-dependent friction) plus randomization of battery voltage, command delay, and ±1° backlash per joint
- Every policy shares a 61-dimensional actor observation (48 proprioception + twist, head pose, body pose commands), enabling hot-swap between walk, recover, and trick policies mid-run
- Software is Apache-2.0, but mechanical and electronic design files are not open
- The robot generates a unique audio identity on first wake that persists permanently; it does not speak in a linguistic sense
- Pre-orders opened August 27, 2026, with deliveries targeted before Christmas
- Sim-to-real hinges on a BAM actuator model (voltage control law, back-EMF, Coulomb/Stribeck/load-dependent friction) plus randomization of battery voltage, command delay, and ±1° backlash per joint
- Every policy shares a 61-dimensional actor observation (48 proprioception + twist, head pose, body pose commands), enabling hot-swap between walk, recover, and trick policies mid-run
- Software is Apache-2.0, but mechanical and electronic design files are not open
- The robot generates a unique audio identity on first wake that persists permanently; it does not speak in a linguistic sense
- Pre-orders opened August 27, 2026, with deliveries targeted before Christmas
Iván Palomares Carrascosa writes about methods for interpreting the dense numerical vector representations, or embeddings, generated by large language models (LLMs). By using a combination of probing classifiers like logistic regression, UMAP dimensionality reduction for visualization, and SHAP values to identify influential latent dimensions, one can analyze the quality and semantic structure captured within LLM-generated embedding spaces.
- Probing classifiers help determine if embeddings are rich enough to distinguish between classes by testing them with simpler models.
- UMAP is used to project high-dimensional embeddings into 2D space for visual inspection of natural groupings.
- SHAP values can pinpoint which specific dimensions in an embedding most significantly influence a classifier's decisions.
- The article demonstrates using Scikit-LLM alongside local Ollama models to generate embeddings cost-effectively.
- Probing classifiers help determine if embeddings are rich enough to distinguish between classes by testing them with simpler models.
- UMAP is used to project high-dimensional embeddings into 2D space for visual inspection of natural groupings.
- SHAP values can pinpoint which specific dimensions in an embedding most significantly influence a classifier's decisions.
- The article demonstrates using Scikit-LLM alongside local Ollama models to generate embeddings cost-effectively.
Chris Patrick writes that SLAC researchers have built a neural-network method for compressing large scientific datasets while preserving fine details that conventional compression erases. The approach uses wavelet analysis to separate data features by scale, then encodes each scale separately through a neural network, enabling 10- to 100-fold file size reductions and selective decompression of only the regions a researcher needs.
- Published in Nature Machine Intelligence (August 24, 2026)
- Motivated by upcoming LCLS upgrades that will generate nearly one terabyte of data per second
- Tested successfully on X-ray diffraction data, solar magnetic field measurements, and photographs
- Neural networks trained on Perlmutter at NERSC (Lawrence Berkeley National Lab)
- Co-developers include researchers from UC Davis and Carnegie Mellon University
- Published in Nature Machine Intelligence (August 24, 2026)
- Motivated by upcoming LCLS upgrades that will generate nearly one terabyte of data per second
- Tested successfully on X-ray diffraction data, solar magnetic field measurements, and photographs
- Neural networks trained on Perlmutter at NERSC (Lawrence Berkeley National Lab)
- Co-developers include researchers from UC Davis and Carnegie Mellon University
El Assadi et al. compare ten LLMs (six families) and 26 embedding models (118M - 14B parameters) on 37 tasks, considering cost. In aggregate, the two paradigms are effectively tied (best LLM scores 77.6 versus best embedding model 77.2), yet their strengths diverge by task: LLMs lead on reasoning-heavy retrieval while embedding models lead on classification, and the two match on clustering, STS, and pair classification.
LLMs are significantly more expensive (up to 1,431x) and slower (2.5-736x) than embedding models for certain tasks. The authors suggest using embedding models for similarity, classification, and clustering, and LLMs for reasoning in retrieval.
Reasoning tokens are 28-81% of LLM inference cost; lower budgets maintain or boost retrieval quality for most tested models.
- Only Gemini 3.1 Pro breaks into the Pareto frontier alongside the leading embedding models.
- Accepted to COLM 2026; code, datasets, and results are publicly released on GitHub.
LLMs are significantly more expensive (up to 1,431x) and slower (2.5-736x) than embedding models for certain tasks. The authors suggest using embedding models for similarity, classification, and clustering, and LLMs for reasoning in retrieval.
Reasoning tokens are 28-81% of LLM inference cost; lower budgets maintain or boost retrieval quality for most tested models.
- Only Gemini 3.1 Pro breaks into the Pareto frontier alongside the leading embedding models.
- Accepted to COLM 2026; code, datasets, and results are publicly released on GitHub.
Alibaba has open-sourced Qwen-UI-Agent, a GUI agent foundation model that operates across mobile, desktop, web, and deep-search environments on real hardware rather than relying on simulation. It achieves top benchmark results: 82.1% on MobileWorld, 79.5% on OSWorld-Verified, and first on WebArena. It also introduces MobileWorld-Real, a 400+ task benchmark on 100+ phones and 150+ apps, with a 92.2% success rate.
- Supports command-line execution alongside standard GUI operations and batches multiple actions into a single decision step to shorten trajectories.
- Built-in safety layer refuses illegal or high-risk requests outright and pauses at sensitive operations (payments, data deletion, privacy grants) for explicit user confirmation.
- Trained via online reinforcement learning on trajectories exceeding 100 steps, paired with adaptive curriculum learning to progressively tackle longer tasks.
- Supports command-line execution alongside standard GUI operations and batches multiple actions into a single decision step to shorten trajectories.
- Built-in safety layer refuses illegal or high-risk requests outright and pauses at sensitive operations (payments, data deletion, privacy grants) for explicit user confirmation.
- Trained via online reinforcement learning on trajectories exceeding 100 steps, paired with adaptive curriculum learning to progressively tackle longer tasks.
occlupanid data writes that the Holotypic Occlupanid Research Group hosts several years of research classifying occlupanids, small ubiquitous objects dotting supermarket aisles and sidewalks, as the most common yet puzzling member of phylum Plasticae within a synthetic taxonomy database.
- The site catalogs dozens of families such as Acutignathidae, Archignathidae, Corrugatidae and Toxodentidae with individual species pages.
- Navigation includes Identification Guide, Publications and Reports, Cartonalia: The Occlupanopsida, and a Guide to symbols for ecological, geographical and taxonomic classification.
- The project also covers morphology, growth and development, origins of the Occlupanida, history of occlupanology, and a Pseudo-occlupanids section.
- The site catalogs dozens of families such as Acutignathidae, Archignathidae, Corrugatidae and Toxodentidae with individual species pages.
- Navigation includes Identification Guide, Publications and Reports, Cartonalia: The Occlupanopsida, and a Guide to symbols for ecological, geographical and taxonomic classification.
- The project also covers morphology, growth and development, origins of the Occlupanida, history of occlupanology, and a Pseudo-occlupanids section.
Samuel K. Ainsworth writes with Jonathan Hayase and Siddhartha Srinivasa that neural network loss landscapes often contain nearly a single basin after accounting for permutation symmetries of hidden units, and they introduce three algorithms to permute one independently trained model into alignment with a reference model so the two can be merged in weight space with near-convex interpolation. Experiments across MLPs, CNNs and ResNets on MNIST, CIFAR-10/100 and ImageNet show the single-basin phenomenon, including the first zero-barrier linear mode connectivity between independently trained ResNets on CIFAR-10, and reveal how mode connectivity emerges with training time and model width. The authors also provide a counterexample showing linear mode connectivity is not universal and discuss limits of the hypothesis.
- Permutation symmetries for a 3-layer 512-width MLP are ~10^3498, far exceeding atoms in the observable universe.
- Weight matching uses coordinate descent on the sum of bilinear assignments problem and runs in seconds without accessing training data.
- Activation matching reduces to a linear assignment problem per layer via ordinary least squares regression on unit activations.
- Straight-through estimator learning of permutations gives the best barriers but is much more expensive than weight or activation matching.
- Merging models trained on disjoint CIFAR-100 splits with weight matching improves test loss and calibration over both inputs while using the compute/memory of one model.
- Permutation symmetries for a 3-layer 512-width MLP are ~10^3498, far exceeding atoms in the observable universe.
- Weight matching uses coordinate descent on the sum of bilinear assignments problem and runs in seconds without accessing training data.
- Activation matching reduces to a linear assignment problem per layer via ordinary least squares regression on unit activations.
- Straight-through estimator learning of permutations gives the best barriers but is much more expensive than weight or activation matching.
- Merging models trained on disjoint CIFAR-100 splits with weight matching improves test loss and calibration over both inputs while using the compute/memory of one model.
Firecrawl introduces pdf-inspector, a high-performance Rust library designed for rapid PDF classification, text extraction, and Markdown conversion. By sampling content streams to quickly distinguish between text-based and scanned documents, the tool enables intelligent routing that bypasses costly OCR services for standard PDFs. It delivers position-aware text extraction, automated table and column detection, and robust encoding handling while maintaining a lightweight footprint with no external ML dependencies or model training requirements.
- Provides bindings for Python, Node.js, and browser WebAssembly environments.
- Achieves sub-200ms processing times on large corpora while outperforming several established local parsers in reading order and table accuracy.
- Features per-page OCR routing suggestions to optimize mixed-format document workflows.
- Handles complex layouts including RTL text, multi-column newspapers, and CID-encoded fonts.
- Released under the MIT license with active community contributions and CI/CD automation.
- Provides bindings for Python, Node.js, and browser WebAssembly environments.
- Achieves sub-200ms processing times on large corpora while outperforming several established local parsers in reading order and table accuracy.
- Features per-page OCR routing suggestions to optimize mixed-format document workflows.
- Handles complex layouts including RTL text, multi-column newspapers, and CID-encoded fonts.
- Released under the MIT license with active community contributions and CI/CD automation.
Rupert Lane writes about Arthur Samuel's checkers program running on the WAITS operating system. Samuel began developing his checkers program in the late 1940s and 1950s, introducing machine learning techniques like rote learning to the field. The article details the program's history from the IBM 7090 to its porting to the PDP-10 at SAIL, where the 1972 version runs on WAITS. It also covers Samuel's work on a home computer version for the VideoBrain and provides instructions for playing the game on WAITS, including board visualization and command syntax.
- The 1952 checkers program was written in raw machine code as the IBM system lacked an assembler at the time.
- Samuel introduced rote learning to the program on the IBM 704 in 1954, enabling interesting gameplay.
- A 1976 match against Duke University was abandoned due to excessive CPU time, though analysis suggests a draw was possible.
- The VideoBrain home console version operated within 1KB of RAM and offered four difficulty levels.
- Users can visualize the board on WAITS with the `BD` command or observe the AI declare wins based on search depths.
- Oldest surviving source file `CKP2.MAC 2,ALS » ` dates to 1971, while the running binary is from 1972.
- The 1952 checkers program was written in raw machine code as the IBM system lacked an assembler at the time.
- Samuel introduced rote learning to the program on the IBM 704 in 1954, enabling interesting gameplay.
- A 1976 match against Duke University was abandoned due to excessive CPU time, though analysis suggests a draw was possible.
- The VideoBrain home console version operated within 1KB of RAM and offered four difficulty levels.
- Users can visualize the board on WAITS with the `BD` command or observe the AI declare wins based on search depths.
- Oldest surviving source file `CKP2.MAC 2,ALS » ` dates to 1971, while the running binary is from 1972.
Supervision provides essential building blocks for computer vision applications, such as data loading and real-time zone counting. The toolkit remains model agnostic, enabling easy integration of various machine learning models via specialized connectors.
- Supports multiple dataset formats including YOLO, COCO, and Pascal VOC
- Offers utilities to split, merge, and convert datasets
- Includes capabilities for speed estimation and dwell time analysis
- Supports multiple dataset formats including YOLO, COCO, and Pascal VOC
- Offers utilities to split, merge, and convert datasets
- Includes capabilities for speed estimation and dwell time analysis
