Category: News

Splitwise: Collaborative Edge-Cloud Inference for LLMs via Lyapunov-Assisted DRLcs.AI updates on arXiv.org arXiv:2512.23310v1 Announce Type: cross
Abstract: Deploying large language models (LLMs) on edge devices is challenging due to their limited memory and power resources. Cloud-only inference reduces device burden but introduces high latency and cost. Static edge-cloud partitions optimize a single metric and struggle when bandwidth fluctuates. We propose Splitwise, a novel Lyapunov-assisted deep reinforcement learning (DRL) framework for fine-grained, adaptive partitioning of LLMs across edge and cloud environments. Splitwise decomposes transformer layers into attention heads and feed-forward sub-blocks, exposing more partition choices than layer-wise schemes. A hierarchical DRL policy, guided by Lyapunov optimization, jointly minimizes latency, energy consumption, and accuracy degradation while guaranteeing queue stability under stochastic workloads and variable network bandwidth. Splitwise also guarantees robustness via partition checkpoints with exponential backoff recovery in case of communication failures. Experiments on Jetson Orin NX, Galaxy S23, and Raspberry Pi 5 with GPT-2 (1.5B), LLaMA-7B, and LLaMA-13B show that Splitwise reduces end-to-end latency by 1.4x-2.8x and cuts energy consumption by up to 41% compared with existing partitioners. It lowers the 95th-percentile latency by 53-61% relative to cloud-only execution, while maintaining accuracy and modest memory requirements.

 arXiv:2512.23310v1 Announce Type: cross
Abstract: Deploying large language models (LLMs) on edge devices is challenging due to their limited memory and power resources. Cloud-only inference reduces device burden but introduces high latency and cost. Static edge-cloud partitions optimize a single metric and struggle when bandwidth fluctuates. We propose Splitwise, a novel Lyapunov-assisted deep reinforcement learning (DRL) framework for fine-grained, adaptive partitioning of LLMs across edge and cloud environments. Splitwise decomposes transformer layers into attention heads and feed-forward sub-blocks, exposing more partition choices than layer-wise schemes. A hierarchical DRL policy, guided by Lyapunov optimization, jointly minimizes latency, energy consumption, and accuracy degradation while guaranteeing queue stability under stochastic workloads and variable network bandwidth. Splitwise also guarantees robustness via partition checkpoints with exponential backoff recovery in case of communication failures. Experiments on Jetson Orin NX, Galaxy S23, and Raspberry Pi 5 with GPT-2 (1.5B), LLaMA-7B, and LLaMA-13B show that Splitwise reduces end-to-end latency by 1.4x-2.8x and cuts energy consumption by up to 41% compared with existing partitioners. It lowers the 95th-percentile latency by 53-61% relative to cloud-only execution, while maintaining accuracy and modest memory requirements. Read More  

Eliminating Inductive Bias in Reward Models with Information-Theoretic Guidancecs.AI updates on arXiv.org arXiv:2512.23461v1 Announce Type: cross
Abstract: Reward models (RMs) are essential in reinforcement learning from human feedback (RLHF) to align large language models (LLMs) with human values. However, RM training data is commonly recognized as low-quality, containing inductive biases that can easily lead to overfitting and reward hacking. For example, more detailed and comprehensive responses are usually human-preferred but with more words, leading response length to become one of the inevitable inductive biases. A limited number of prior RM debiasing approaches either target a single specific type of bias or model the problem with only simple linear correlations, textit{e.g.}, Pearson coefficients. To mitigate more complex and diverse inductive biases in reward modeling, we introduce a novel information-theoretic debiasing method called textbf{D}ebiasing via textbf{I}nformation optimization for textbf{R}M (DIR). Inspired by the information bottleneck (IB), we maximize the mutual information (MI) between RM scores and human preference pairs, while minimizing the MI between RM outputs and biased attributes of preference inputs. With theoretical justification from information theory, DIR can handle more sophisticated types of biases with non-linear correlations, broadly extending the real-world application scenarios for RM debiasing methods. In experiments, we verify the effectiveness of DIR with three types of inductive biases: textit{response length}, textit{sycophancy}, and textit{format}. We discover that DIR not only effectively mitigates target inductive biases but also enhances RLHF performance across diverse benchmarks, yielding better generalization abilities. The code and training recipes are available at https://github.com/Qwen-Applications/DIR.

 arXiv:2512.23461v1 Announce Type: cross
Abstract: Reward models (RMs) are essential in reinforcement learning from human feedback (RLHF) to align large language models (LLMs) with human values. However, RM training data is commonly recognized as low-quality, containing inductive biases that can easily lead to overfitting and reward hacking. For example, more detailed and comprehensive responses are usually human-preferred but with more words, leading response length to become one of the inevitable inductive biases. A limited number of prior RM debiasing approaches either target a single specific type of bias or model the problem with only simple linear correlations, textit{e.g.}, Pearson coefficients. To mitigate more complex and diverse inductive biases in reward modeling, we introduce a novel information-theoretic debiasing method called textbf{D}ebiasing via textbf{I}nformation optimization for textbf{R}M (DIR). Inspired by the information bottleneck (IB), we maximize the mutual information (MI) between RM scores and human preference pairs, while minimizing the MI between RM outputs and biased attributes of preference inputs. With theoretical justification from information theory, DIR can handle more sophisticated types of biases with non-linear correlations, broadly extending the real-world application scenarios for RM debiasing methods. In experiments, we verify the effectiveness of DIR with three types of inductive biases: textit{response length}, textit{sycophancy}, and textit{format}. We discover that DIR not only effectively mitigates target inductive biases but also enhances RLHF performance across diverse benchmarks, yielding better generalization abilities. The code and training recipes are available at https://github.com/Qwen-Applications/DIR. Read More  

Timepoint-Specific Benchmarking of Deep Learning Models for Glioblastoma Follow-Up MRIcs.AI updates on arXiv.org arXiv:2511.18595v2 Announce Type: replace-cross
Abstract: Differentiating true tumor progression (TP) from treatment-related pseudoprogression (PsP) in glioblastoma remains challenging, especially at early follow-up. We present the first stage-specific, cross-sectional benchmarking of deep learning models for follow-up MRI using the Burdenko GBM Progression cohort (n = 180). We analyze different post-RT scans independently to test whether architecture performance depends on time-point. Eleven representative DL families (CNNs, LSTMs, hybrids, transformers, and selective state-space models) were trained under a unified, QC-driven pipeline with patient-level cross-validation. Across both stages, accuracies were comparable (~0.70-0.74), but discrimination improved at the second follow-up, with F1 and AUC increasing for several models, indicating richer separability later in the care pathway. A Mamba+CNN hybrid consistently offered the best accuracy-efficiency trade-off, while transformer variants delivered competitive AUCs at substantially higher computational cost and lightweight CNNs were efficient but less reliable. Performance also showed sensitivity to batch size, underscoring the need for standardized training protocols. Notably, absolute discrimination remained modest overall, reflecting the intrinsic difficulty of TP vs. PsP and the dataset’s size imbalance. These results establish a stage-aware benchmark and motivate future work incorporating longitudinal modeling, multi-sequence MRI, and larger multi-center cohorts.

 arXiv:2511.18595v2 Announce Type: replace-cross
Abstract: Differentiating true tumor progression (TP) from treatment-related pseudoprogression (PsP) in glioblastoma remains challenging, especially at early follow-up. We present the first stage-specific, cross-sectional benchmarking of deep learning models for follow-up MRI using the Burdenko GBM Progression cohort (n = 180). We analyze different post-RT scans independently to test whether architecture performance depends on time-point. Eleven representative DL families (CNNs, LSTMs, hybrids, transformers, and selective state-space models) were trained under a unified, QC-driven pipeline with patient-level cross-validation. Across both stages, accuracies were comparable (~0.70-0.74), but discrimination improved at the second follow-up, with F1 and AUC increasing for several models, indicating richer separability later in the care pathway. A Mamba+CNN hybrid consistently offered the best accuracy-efficiency trade-off, while transformer variants delivered competitive AUCs at substantially higher computational cost and lightweight CNNs were efficient but less reliable. Performance also showed sensitivity to batch size, underscoring the need for standardized training protocols. Notably, absolute discrimination remained modest overall, reflecting the intrinsic difficulty of TP vs. PsP and the dataset’s size imbalance. These results establish a stage-aware benchmark and motivate future work incorporating longitudinal modeling, multi-sequence MRI, and larger multi-center cohorts. Read More  

How Cloudflare’s tokio-quiche Makes QUIC and HTTP/3 a First Class Citizen in Rust BackendsMarkTechPost Cloudflare has open sourced tokio-quiche, an asynchronous QUIC and HTTP/3 Rust library that wraps its battle tested quiche implementation with the Tokio runtime. The library has been refined inside production systems such as Apple iCloud Private Relay, next generation Oxy based proxies and WARP’s MASQUE client, where it handles millions of HTTP/3 requests per second
The post How Cloudflare’s tokio-quiche Makes QUIC and HTTP/3 a First Class Citizen in Rust Backends appeared first on MarkTechPost.

 Cloudflare has open sourced tokio-quiche, an asynchronous QUIC and HTTP/3 Rust library that wraps its battle tested quiche implementation with the Tokio runtime. The library has been refined inside production systems such as Apple iCloud Private Relay, next generation Oxy based proxies and WARP’s MASQUE client, where it handles millions of HTTP/3 requests per second
The post How Cloudflare’s tokio-quiche Makes QUIC and HTTP/3 a First Class Citizen in Rust Backends appeared first on MarkTechPost. Read More  

Production-Ready LLMs Made Simple with the NeMo Agent ToolkitTowards Data Science From simple chat to multi-agent reasoning and real-time REST APIs
The post Production-Ready LLMs Made Simple with the NeMo Agent Toolkit appeared first on Towards Data Science.

 From simple chat to multi-agent reasoning and real-time REST APIs
The post Production-Ready LLMs Made Simple with the NeMo Agent Toolkit appeared first on Towards Data Science. Read More  

How to Design Transactional Agentic AI Systems with LangGraph Using Two-Phase Commit, Human Interrupts, and Safe RollbacksMarkTechPost In this tutorial, we implement an agentic AI pattern using LangGraph that treats reasoning and action as a transactional workflow rather than a single-shot decision. We model a two-phase commit system in which an agent stages reversible changes, validates strict invariants, pauses for human approval via graph interrupts, and commits or rolls back only then.
The post How to Design Transactional Agentic AI Systems with LangGraph Using Two-Phase Commit, Human Interrupts, and Safe Rollbacks appeared first on MarkTechPost.

 In this tutorial, we implement an agentic AI pattern using LangGraph that treats reasoning and action as a transactional workflow rather than a single-shot decision. We model a two-phase commit system in which an agent stages reversible changes, validates strict invariants, pauses for human approval via graph interrupts, and commits or rolls back only then.
The post How to Design Transactional Agentic AI Systems with LangGraph Using Two-Phase Commit, Human Interrupts, and Safe Rollbacks appeared first on MarkTechPost. Read More