January 2026

Generating realistic graph-structured data is challenging due to discrete structures, variable sizes, and class-specific connectivity patterns that resist conventional generative modelling. While recent graph generation methods employ generative adversarial network (GAN) frameworks to handle permutation invariance and irregular topologies, they typically rely on random edge sampling with fixed probabilities, limiting their capacity to capture complex structural dependencies between nodes. We propose a density-aware conditional graph generation framework using Wasserstein GANs (WGAN) that replaces random sampling with a learnable […]

Ver mais

Mobile vision care goes global OptikosPrime just obliterated the primary barrier in digital optometry. The company announced today that it has successfully derived a complete eye prescription using only a standard smartphone camera. Developed under the internal platform name Argus, its vision tests bypass the need for traditional optometry hardware or specialized clinical personnel. In short, OptikosPrime’s Argus provides a prescription based on a selfie. The global eyewear market is projected to reach $192.74 billion in 2026, according […]

Ver mais

We built a WASM-based sandbox for running LLM-generated code in agentic workflows. The problem: most agent frameworks execute code via subprocess or exec() directly on the host. One prompt injection and you’re exposed. Our approach: QuickJS runtime compiled to WASM (no syscalls, no network, no filesystem escape) Capability-based tool access—agents can only call functions you explicitly provide Per-tool constraints (e.g., Param(“amount”) <= 1000) Virtual filesystem that resets between executions It’s a Python package wrapping a Rust/WASM binary. Install […]

Ver mais

We study a class of iterated empirical risk minimization (ERM) procedures in which two successive ERMs are performed on the same dataset, and the predictions of the first estimator enter as an argument in the loss function of the second. This setting, which arises naturally in active learning and reweighting schemes, introduces intricate statistical dependencies across samples and fundamentally distinguishes the problem from classical single-stage ERM analyses. For linear models trained with a broad class of convex losses […]

Ver mais

Offline Reinforcement Learning (RL) relies on policy constraints to mitigate extrapolation error, where both the constraint form and constraint strength critically shape performance. However, most existing methods commit to a single constraint family: weighted behavior cloning, density regularization, or support constraints, without a unified principle that explains their connections or trade-offs. In this work, we propose Continuous Constraint Interpolation (CCI), a unified optimization framework in which these three constraint families arise as special cases along a common constraint […]

Ver mais

We explore the Traveling Salesman Problem (TSP) using a Variational Quantum Algorithm (VQA), with a focus on representation efficiency and model structure learning rather than just parameter tuning. Key ideas: Compact permutation-based encoding Uses O(nlog⁡n)O(n log n)O(nlogn) qubits and guarantees that every quantum state corresponds to a valid tour (no constraint penalties or repair steps). Adaptive circuit optimization Instead of fixing the quantum circuit (ansatz) upfront, we optimize its structure using Simulated Annealing: add / remove rotation and […]

Ver mais

Hi everyone, I’m working on a university project and could really use some advice from people with more experience in autonomous navigation / RL / simulation. Task: I need to design a deep learning model that directly controls a drone (x, y, z, pitch, yaw — roll probably doesn’t make much sense here 😅). The drone should autonomously patrol and map indoor and outdoor environments. Example use case: A warehouse where the drone automatically flies through all aisles […]

Ver mais

I’m excited to share a major update to sklearn-diagnose – the open-source Python library that acts as an “MRI scanner” for your ML models (https://www.reddit.com/r/MachineLearning/s/EcMRYPVIDX) When I first released sklearn-diagnose, users could generate diagnostic reports to understand why their models were failing. But I kept thinking – what if you could talk to your diagnosis? What if you could ask follow-up questions and drill down into specific issues? Now you can! 🚀 🆕 What’s New: Interactive Diagnostic Chatbot […]

Ver mais

A question for people working in RL and image generative models (diffusion, flow based etc). There seems to be more emerging work in RL fine tuning techniques for these models (e.g. DDPO, DiffusionNFT, etc). I’m interested to know – is it crazy to try to train these models from scratch with a reward signal only (i.e without any supervision data from a random initialised policy)? And specifically, what techniques could be used to overcome issues with reward sparsity […]

Ver mais