4arXiv cs.LG (Machine Learning)·
Blueprint for thermodynamic computing stack using stochastic analog superconducting circuits for energy-efficient ML
Researchers propose a blueprint for a thermodynamic computing stack that uses Langevin dynamics with tunable energy potentials implemented in stochastic analog hardware to run probabilistic machine learning workloads. The framework maps popular ML model classes onto hardware-native energy-based models via probabilistic graphical models, with runtime and energy analysis via theory and simulation. A preliminary experimental realization using superconducting circuits driven by thermal noise is presented. The work targets the energy and latency bottlenecks of conventional ML hardware.
Related guides (2)
Related events (8)
Fault-tolerant LLM pipeline achieves autonomous end-to-end research in computational physics
Researchers present a multi-phase LLM pipeline that autonomously processes 11,083 condensed-matter physics arXiv papers, conceives a research direction, calibrates methodology against published references, runs first-principles computations, and produces a publication-grade manuscript on altermagnetic piezomagnetism. The system operates across 47 fresh-context sessions with 2,162 literature-consultation events, using redundancy and adversarial review for fault tolerance, with human intervention limited to operational knowledge curation at reproduction failures rather than scientific direction. Ablation experiments isolate 'structurally enforced numerical confrontation' at calibration checkpoints as the key mechanism preventing hallucination in physical science domains. The work extends autonomous research agents beyond ML sandboxes into high-stakes physical science, where external literature anchors replace execution-based calibration.
LatentFlow: Training-free framework for conditioning stochastic processes via latent-space inference
LatentFlow is a new general framework for conditioning stochastic processes without learned neural approximations or training. The approach maps a stochastic process to a tractable latent innovation space, reduces conditioning to latent-space inference via a guided probability flow, and pushes samples forward — yielding provably exact conditioning at the level of the target law. The method is claimed to work across a broad range of model classes (spatial priors, stochastic PDEs, mechanistic models, point processes) and runs in seconds on a single CPU.
Implicit Generation and Generalization Methods for Energy-Based Models
OpenAI published research on stable and scalable training of energy-based models (EBMs), achieving sample quality competitive with GANs at low temperatures while retaining mode coverage guarantees of likelihood-based models. The approach uses iterative compute during generation to continually refine outputs. This work positions EBMs as a promising alternative generative modeling paradigm bridging GANs and likelihood-based models.
Theoretical framework proposes native meta-architecture to replace application-layer LLM cognitive simulation
A preprint from arXiv proposes a theoretical framework for embedding cognitive architecture natively into LLMs rather than simulating it via prompt engineering and context management. The framework introduces three mechanisms: Structural Tension (an endogenous loss function from information-manifold conflict), an Offline Recurrent Loop (sandboxed self-processing without external input), and Inference-time Plasticity (topology reconfiguration without weight modification). The authors argue these mechanisms could produce heterogeneous model instances with distinct topological structures through path-dependent evolution, while remaining within governance rails. The paper is primarily theoretical, offering operational definitions, reconfiguration operators, and falsification criteria rather than empirical results.
Accelerating Stable Diffusion Inference on Intel CPUs
This Hugging Face blog post details techniques for optimizing Stable Diffusion inference on Intel CPUs, likely covering quantization, operator fusion, and Intel-specific hardware acceleration libraries. The post addresses the practical challenge of running diffusion models on CPU hardware without dedicated GPUs. This is relevant to inference economics and enterprise deployment patterns where GPU availability is constrained.
MLIR-based compilation method for LLM inference on specialized hardware
Researchers present an MLIR-based compiler pipeline for deploying large language models on AI accelerators, using two dialect layers (TopOp for framework-agnostic graph representation and TpuOp for hardware-specific lowering). The method splits each Transformer layer into three static compilation stages (prefill, prefill_kv, decode) to handle the distinct computational profiles of prompt processing and autoregressive generation. The approach is implemented in the open-source TPU-MLIR compiler and LLM-TPU project, supporting Qwen, Llama, InternVL, and MiniCPM-V families with GPTQ, AWQ, and AutoRound quantization.
Fine-tuning Stable Diffusion models on Intel CPUs
This Hugging Face blog post describes a workflow for fine-tuning Stable Diffusion image generation models on Intel CPUs rather than GPUs. It covers the tooling and optimizations required to make CPU-based diffusion model training practical, relevant to inference-economics and hardware diversification trends. The post targets practitioners looking to reduce dependency on GPU hardware for generative model fine-tuning.
PALS: Power-Aware LLM Serving Runtime for MoE and Dense Models
PALS is a power-aware inference runtime integrated into vLLM that treats GPU power caps as a first-class scheduling parameter alongside batch size and parallelism settings. Using lightweight offline power-performance models and a feedback-driven controller, it jointly optimizes energy efficiency and throughput targets without model retraining or API changes. Across multi-GPU deployments with both dense and MoE models, PALS achieves up to 26.3% energy efficiency improvement and reduces QoS violations by 4-7x under power constraints, enabling energy-proportional and grid-interactive AI serving.