#Computational Efficiency
Latest news articles tagged with "Computational Efficiency". Follow the timeline of events, related topics, and entities.
Articles (23)
-
🇺🇸 Mitigating Legibility Tax with Decoupled Prover-Verifier Games
[USA]
arXiv:2602.23248v1 Announce Type: new Abstract: As large language models become increasingly capable, it is critical that their outputs can be easily checked by less capable systems. Prover-verifier ...
Related: #Artificial Intelligence, #Model Verification -
🇺🇸 S2O: Early Stopping for Sparse Attention via Online Permutation
[USA]
arXiv:2602.22575v1 Announce Type: cross Abstract: Attention scales quadratically with sequence length, fundamentally limiting long-context inference. Existing block-granularity sparsification can red...
Related: #Machine Learning Optimization, #Attention Mechanisms -
🇺🇸 ArchAgent: Agentic AI-driven Computer Architecture Discovery
[USA]
arXiv:2602.22425v1 Announce Type: new Abstract: Agile hardware design flows are a critically needed force multiplier to meet the exploding demand for compute. Recently, agentic generative AI systems ...
Related: #AI in Hardware Design, #Automated Architecture Discovery -
🇺🇸 ECHO: Encoding Communities via High-order Operators
[USA]
arXiv:2602.22446v1 Announce Type: cross Abstract: Community detection in attributed networks faces a fundamental divide: topological algorithms ignore semantic features, while Graph Neural Networks (...
Related: #Machine Learning Innovation, #Network Analysis -
🇺🇸 From Shallow Bayesian Neural Networks to Gaussian Processes: General Convergence, Identifiability and Scalable Inference
[USA]
arXiv:2602.22492v1 Announce Type: cross Abstract: In this work, we study scaling limits of shallow Bayesian neural networks (BNNs) via their connection to Gaussian processes (GPs), with an emphasis o...
Related: #Machine Learning Theory, #Statistical Modeling, #Neural Networks -
🇺🇸 Stable Adaptive Thinking via Advantage Shaping and Length-Aware Gradient Regulation
[USA]
arXiv:2602.22556v1 Announce Type: cross Abstract: Large reasoning models (LRMs) achieve strong performance through extended reasoning traces, but they often exhibit overthinking behavior for low-comp...
Related: #Machine Learning, #Artificial Intelligence -
🇺🇸 CHESS: Context-aware Hierarchical Efficient Semantic Selection for Long-Context LLM Inference
[USA]
arXiv:2602.20732v1 Announce Type: new Abstract: Long-context LLMs demand accurate inference at low latency, yet decoding becomes primarily constrained by KV cache as context grows. Prior pruning meth...
Related: #AI Optimization, #Large Language Models -
🇺🇸 HELP: HyperNode Expansion and Logical Path-Guided Evidence Localization for Accurate and Efficient GraphRAG
[USA]
arXiv:2602.20926v1 Announce Type: new Abstract: Large Language Models (LLMs) often struggle with inherent knowledge boundaries and hallucinations, limiting their reliability in knowledge-intensive ta...
Related: #Artificial Intelligence, #Knowledge Retrieval -
🇺🇸 MoBiQuant: Mixture-of-Bits Quantization for Token-Adaptive Elastic LLMs
[USA]
arXiv:2602.20191v1 Announce Type: cross Abstract: Changing runtime complexity on cloud and edge devices necessitates elastic large language model (LLM) deployment, where an LLM can be inferred with v...
Related: #Machine Learning, #Quantization Optimization -
🇺🇸 TIFO: Time-Invariant Frequency Operator for Stationarity-Aware Representation Learning in Time Series
[USA]
arXiv:2602.17122v1 Announce Type: cross Abstract: Nonstationary time series forecasting suffers from the distribution shift issue due to the different distributions that produce the training and test...
Related: #Time‑Series Forecasting, #Frequency Domain Analysis, #Stationarity vs. Non‑Stationarity, #Distribution Shift Mitigation -
🇺🇸 Beyond Message Passing: A Symbolic Alternative for Expressive and Interpretable Graph Learning
[USA]
arXiv:2602.16947v1 Announce Type: cross Abstract: Graph Neural Networks (GNNs) have become essential in high-stakes domains such as drug discovery, yet their black-box nature remains a significant ba...
Related: #Graph Neural Networks, #Symbolic Machine Learning, #Expressivity Limits, #Interpretability -
🇺🇸 SubQuad: Near-Quadratic-Free Structure Inference with Distribution-Balanced Objectives in Adaptive Receptor framework
[USA]
arXiv:2602.17330v1 Announce Type: cross Abstract: Comparative analysis of adaptive immune repertoires at population scale is hampered by two practical bottlenecks: the near-quadratic cost of pairwise...
Related: #Machine Learning, #Adaptive Immune Repertoire Analysis, #Bias Mitigation, #Clustering Algorithms -
🇺🇸 Accelerating Large-Scale Dataset Distillation via Exploration-Exploitation Optimization
[USA]
arXiv:2602.15277v1 Announce Type: cross Abstract: Dataset distillation compresses the original data into compact synthetic datasets, reducing training time and storage while retaining model performan...
Related: #Dataset Distillation, #Large‑Scale Machine Learning, #Optimization Techniques, #Exploration vs. Exploitation -
🇺🇸 Rational Neural Networks have Expressivity Advantages
[USA]
arXiv:2602.12390v1 Announce Type: cross Abstract: We study neural networks with trainable low-degree rational activation functions and show that they are more expressive and parameter-efficient than ...
Related: #Neural Network Architecture, #Theoretical Computer Science -
🇺🇸 SLA2: Sparse-Linear Attention with Learnable Routing and QAT
[USA]
arXiv:2602.12675v1 Announce Type: cross Abstract: Sparse-Linear Attention (SLA) combines sparse and linear attention to accelerate diffusion models and has shown strong performance in video generatio...
Related: #Machine Learning, #Attention Mechanisms -
🇺🇸 Unleashing Low-Bit Inference on Ascend NPUs: A Comprehensive Evaluation of HiFloat Formats
[USA]
arXiv:2602.12635v1 Announce Type: cross Abstract: As LLMs scale, low-bit floating-point formats like MXFP and NVFP4 offer new opportunities for precision and efficiency. In this work, we evaluate HiF...
Related: #AI Hardware, #Precision Optimization -
🇺🇸 SimpleMatch: A Simple and Strong Baseline for Semantic Correspondence
[USA]
arXiv:2601.12357v2 Announce Type: replace-cross Abstract: Recent advances in semantic correspondence have been largely driven by the use of pre-trained large-scale models. However, a limitation of th...
Related: #Computer Vision, #Artificial Intelligence -
🇺🇸 Hierarchical Retrieval at Scale: Bridging Transparency and Efficiency
[USA]
arXiv:2502.07971v2 Announce Type: replace-cross Abstract: Information retrieval is a core component of many intelligent systems as it enables conditioning of outputs on new and large-scale datasets. ...
Related: #Information Retrieval, #Explainable AI -
🇺🇸 Does Your Reasoning Model Implicitly Know When to Stop Thinking?
[USA]
arXiv:2602.08354v1 Announce Type: new Abstract: Recent advancements in large reasoning models (LRMs) have greatly improved their capabilities on complex reasoning tasks through Long Chains of Thought...
Related: #Artificial Intelligence, #Machine Learning -
🇺🇸 CoRefine: Confidence-Guided Self-Refinement for Adaptive Test-Time Compute
[USA]
arXiv:2602.08948v1 Announce Type: new Abstract: Large Language Models (LLMs) often rely on test-time scaling via parallel decoding (for example, 512 samples) to boost reasoning accuracy, but this inc...
Related: #Artificial Intelligence, #Machine Learning -
🇺🇸 Jackpot: Optimal Budgeted Rejection Sampling for Extreme Actor-Policy Mismatch Reinforcement Learning
[USA]
arXiv:2602.06107v1 Announce Type: new Abstract: Reinforcement learning (RL) for large language models (LLMs) remains expensive, particularly because the rollout is expensive. Decoupling rollout gener...
Related: #Artificial Intelligence, #Machine Learning -
🇺🇸 HyPER: Bridging Exploration and Exploitation for Scalable LLM Reasoning with Hypothesis Path Expansion and Reduction
[USA]
arXiv:2602.06527v1 Announce Type: new Abstract: Scaling test-time compute with multi-path chain-of-thought improves reasoning accuracy, but its effectiveness depends critically on the exploration-exp...
Related: #Artificial Intelligence, #Machine Learning -
🇺🇸 CtrlCoT: Dual-Granularity Chain-of-Thought Compression for Controllable Reasoning
[USA]
arXiv:2601.20467v1 Announce Type: new Abstract: Chain-of-thought (CoT) prompting improves LLM reasoning but incurs high latency and memory cost due to verbose traces, motivating CoT compression with ...
Related: #Artificial Intelligence, #Language Models