Home Knowledge Base Budget Predictor

Budget Predictor

No mentions found

This entity hasn't been tracked yet, or Iris is still building its knowledge base.

Related Articles from SNS

BUDDY: BUdget-Driven DYnamic Depth Routing for Adaptive Large Language Model Inference

arXiv:2606.09514v1 Announce Type: new Abstract: Large language models (LLMs) incur high inference cost due to their depth and parameter scale. Depth pruning can reduce latency by skipping redundant Transformer blocks, but existing methods (i) provide limited control under user-specific compute budgets and (ii) typically fix the routing path, failing to adapt as the context grows during decoding. We propose Buddy, a budget-driven dynamic depth routing framework.

arXiv CS 1d ago

IMWM: Intuition Models Complement World Models for Latent Planning

arXiv:2606.01626v1 Announce Type: new Abstract: Planning with a learned latent world model is a promising route to control from raw pixels, but a strong world model alone is not enough. We show this experimentally: even with a perfect world model (operationalized by replacing the learned forward predictor with an idealized rollout of the true environment dynamics), a finite-budget sample-based planner still fails on some tasks, indicating that the bottleneck can lie in search rather than in...

arXiv CS 8d ago

AutoLab: Can Frontier Models Solve Long-Horizon Auto Research and Engineering Tasks?

arXiv:2606.05080v1 Announce Type: new Abstract: Scientific and engineering progress is fundamentally a long-horizon iterative process: proposing changes, running experiments, measuring outcomes, and continuously refining artifacts. Yet existing benchmarks for frontier models primarily evaluate either single-turn responses or short-horizon agent trajectories, failing to capture the challenges of sustained iterative improvement over extended time horizons. To address this gap, we introduce...

arXiv CS 6d ago

Not All Errors Are Equal: Consequence-Aware Reasoning Compute Allocation

Announce Type: new Abstract: Modern reasoning models can allocate different amounts of test-time computation, such as thinking tokens, model calls, or compute budget, to different tasks. Existing methods generally drive this allocation by predicted difficulty and spend more compute where it is expected to raise accuracy. This implicitly assumes that all failures cost the same, since an accuracy objective weights every task equally.

arXiv CS 6d ago

Risk-Aware Planning for Transit Desert Remediation Under Demand Uncertainty

Announce Type: new Abstract: Transit deserts are areas where public transportation is inadequate despite evidence of travel demand, a condition that affects tens of millions of residents across the Americas. Planning for these areas is difficult because the usual demand signal is missing: ridership cannot be observed before service exists. To address that setting, we formulate risk-aware transit desert remediation as a partially observable Markov decision process with Conditional...

arXiv CS 1d ago

PI-JEPA: Label-Free Surrogate Pretraining for Coupled Multiphysics Simulation via Operator-Split Latent Prediction

Announce Type: replace Abstract: Reservoir simulation workflows face a fundamental data asymmetry: input parameter fields (geostatistical permeability realizations, porosity distributions) are free to generate in arbitrary quantities, yet existing neural operator surrogates require large corpora of expensive labeled simulation trajectories and cannot exploit this unlabeled structure. We introduce \textbf{PI-JEPA} (Physics-Informed Joint Embedding Predictive Architecture), a surrogate...

arXiv CS 5d ago

PI-JEPA: Label-Free Surrogate Pretraining for Coupled Multiphysics Simulation via Operator-Split Latent Prediction

Announce Type: replace-cross Abstract: Reservoir simulation workflows face a fundamental data asymmetry: input parameter fields (geostatistical permeability realizations, porosity distributions) are free to generate in arbitrary quantities, yet existing neural operator surrogates require large corpora of expensive labeled simulation trajectories and cannot exploit this unlabeled structure. We introduce \textbf{PI-JEPA} (Physics-Informed Joint Embedding Predictive Architecture), a surrogate...

arXiv Physics 5d ago

Learning-to-Defer in Non-Stationary Time Series via Switching State-Space Models

Announce Type: replace Abstract: Learning-to-defer (L2D) routes each decision to a system's own predictor or to an external expert. Streaming time-series settings break the offline-L2D assumptions: the data are non-stationary, expert availability shifts over time, and the internal predictor is trained online. We propose L2D-SLDS, a one-stage online L2D framework based on a factorized switching linear-Gaussian state-space model over all potential residuals: a discrete regime, a shared global...

arXiv CS 9d ago

Why Everyone Hates AI Data Centers

The Atlantic 12d ago

Victoria will be Labor's first real One Nation electoral test

analysis Even Hanson knows polling isn't everything but major parties have their work cut out Tue 2 Jun 2026 at 5:00am When it comes to reading political polling, it is worth taking a leaf out of Pauline Hanson's book. The One Nation leader, whose popularity has soared in the past 12 months, would be the first to say she doesn't put too much stock in the numbers. That's not to suggest the results aren't to be taken seriously, only that a poll can only reveal so much.

ABC Australia 9d ago