Framework disentangles deep feature uncertainty for efficient inference.
problem Inference-time uncertainty estimation for reliable decision-making.
method Uncertainty-Guided Inference-Time Selection framework.
result Significantly tighter prediction intervals and 60% compute reduction.
New method aligns diffusion models for inference-time properties without retraining.
problem Aligning pre-trained diffusion models for desired inference-time properties.
method Variationally stable Doob's matching for provable guidance estimation.
result Consistent estimator of guidance with non-asymptotic convergence guarantees.
DriftLite improves inference quality of diffusion models without retraining.
problem Adapting pre-trained diffusion models to new target distributions without retraining.
method Lightweight, training-free particle-based approach that steers inference dynamics with optimal stability control.
result Consistently reduces variance and improves sample quality over existing methods.
Tutorial on optimizing diffusion model samples for specific metrics.
problem Optimizing diffusion model samples for specific downstream metrics.
method Review and exploration of inference-time guidance and alignment methods.
result Unified perspective on inference-time algorithms and novel methods.
EHR-MPC optimizes sepsis treatment using digital twins and inference-time control.
problem Optimal sepsis treatment policies are contested and difficult to adapt during inference.
method EHR-MPC decouples learning patient dynamics from treatment optimization, enabling inference-time control over learned digital twins.
result EHR-MPC achieves comparable off-policy performance and improved simulation performance compared to RL baselines.
Improves diffusion model performance and efficiency through classical search.
problem Tackles inference-time control in diffusion models.
method Proposes a framework combining local and global search for efficient navigation.
result Significant gains in performance and efficiency across various domains.
Best-of-∞ improves LLM performance by efficiently allocating inference-time computation.
problem Achieving optimal performance in test-time LLM ensembling with infinite budget.
method Adaptive generation scheme and weighted ensembles of LLMs, formulated as mixed-integer linear program.
result Optimal ensemble weighting improves performance over individual models.
Rate-In dynamically adjusts dropout rates during inference to improve uncertainty estimation in neural networks.
problem Static dropout rates lead to suboptimal uncertainty estimates in neural networks.
method Rate-In dynamically adjusts dropout rates using information-theoretic principles.
result Rate-In improves calibration and sharpens uncertainty estimates compared to fixed or heuristic dropout rates.
DTS improves inference-time alignment of diffusion models with less compute.
problem Inference-time alignment of diffusion models suffers from inaccurate value estimation and inefficient reuse of past computations.
method Diffusion Tree Sampling (DTS) uses a tree-based approach to propagate terminal rewards and iteratively refine value estimates.
result DTS produces asymptotically exact samples and matches the FID of best-performing baselines with up to 10x less compute.
Paper explores BERT's efficiency on SQuAD2.0, freezing layers and using adapters.
problem Improving BERT's efficiency for SQuAD2.0 while maintaining performance.
method Freezing transformer layers, using adapters, and context-aware convolutional filters.
result Context-aware convolutional filters do not improve practical efficiency.
Optimizes trading policies using future price forecasts.
problem Static reinforcement learning agents lack mechanisms for using price forecasts at inference time.
method FPILOT framework inspired by Model Predictive Control (MPC). Uses a predictive model to construct an allocation-based imagined return objective at each decision step.
result Consistent improvements in total return and risk-adjusted metrics across various policy learning algorithms.
Inference-Time Scaling can be extended to domains prone to systematic failure using intrinsic statistics.
problem Scaling inference time in domains prone to systematic failure
method Intrinsic Selection (iS), Intrinsic Particle Filtering (iPF), and Particle Distillation (dPF)
result Intrinsic Selection improves engineering design selection by 20% and pass@1 by 6.1 points on average.
PETAL adapts models to changing target domains over time.
problem Lifelong test-time adaptation in changing target domains.
method Probabilistic framework with student-teacher model and data-driven parameter restoration.
result PETAL achieves better results than state-of-the-art for online lifelong test-time adaptation.
Inference-aware meta-alignment of LLMs reduces computational cost.
problem Aligning LLMs to diverse human preferences is challenging due to conflicting criteria.
method IAMA trains a base model to be aligned to multiple tasks via different inference-time alignment algorithms, using non-linear GRPO for optimization.
result IAMA enables effective alignment of LLMs to multiple criteria with limited computational budget.
RNE provides a flexible framework for diffusion models, enabling inference-time control and energy-based training.
problem Insufficient knowledge of marginal densities in diffusion models.
method Introduces Radon-Nikodym Estimator (RNE) to reveal the connection between marginal densities and transition kernels.
result RNE delivers strong results in inference-time control and energy-based diffusion training.
This work makes neural sequence models more efficient by controlling computation.
problem Fixed compute for all examples in neural networks.
method Conditional computation to adapt compute to example complexity.
result Conditional Computation Transformer (CCT) improves efficiency and performance.
Swift Hydra uses RL and generative AI to improve anomaly detection.
problem Generalization to unseen anomalies in critical systems.
method Generative AI and reinforcement learning (RL) for synthesizing diverse anomaly samples.
result Swift Hydra outperforms state-of-the-art models on ADBench benchmark.
A new method calibrates scientific models by adding randomness to their predictions.
problem Current scientific foundation models lack calibrated uncertainty.
method Stochastic Attention, which randomizes attention weights using multinomial samples.
result Stochastic Attention achieves the strongest native calibration and sharpest prediction intervals.
This work shows how approximate reward models can significantly improve inference-time scaling.
problem Improving the efficiency of inference for large language models.
method Identifying the Bellman error of approximate reward models and using Sequential Monte Carlo (SMC) for inference.
result Approximate reward models can reduce computational complexity from exponential to polynomial in T. New algorithm improves inference-time alignment without reward hacking.
problem Improving quality of responses from language models with limited compute.
method Inference-time alignment, focusing on extttInferenceTimePessimism algorithm. result Optimal performance and scaling-monotonicity of extttInferenceTimePessimism. STAD adapts models to evolving time-based data shifts.
problem Gradual distribution shifts over time challenge existing test-time adaptation methods.
method Bayesian filtering method that learns time-varying dynamics in hidden features.
result STAD excels in handling small batch sizes and label shift on real-world data.
Distributed sensors compress and send features to a fusion center for linear regression.
problem Efficiently compress and transmit features from distributed sensors to a fusion center under varying communication constraints.
method Designs a distributed and adaptive feature compression scheme using optimal quantizers and simple adaptive strategies.
result Demonstrates improved inference performance through simulated experiments.
URGE improves diffusion model quality without gradients or Hessian.
problem Improving sample quality in diffusion models without gradient evaluations.
method Path-wise importance reweighting via Girsanov change of measure.
result URGE achieves better generation quality than existing methods.
End-to-end meta-learned system for image compression.
problem Reducing the gap between training and inference conditions in image compression.
method Model-Agnostic Meta-learning approach for latent tensor overfitting and updating encoder and decoder networks.
result Meta-learned system achieves better compression performance compared to traditional methods.
Paper adapts ACI for online multi-step time-series forecasting with coverage guarantees.
problem Achieving reliable error bounds in online multi-step time-series forecasting.
method Adaptive conformal inference (ACI) adapted for multi-step forecasting with dynamic significance levels.
result Proposes a multi-step ACI algorithm with finite-sample coverage guarantees for non-exchangeable data.
The problem of keyword spotting i.e. identifying keywords in a real-time audio stream is mainly solved by applying a neural network over successive sliding windows. Due to the difficulty of the task, baseline models are usually large, resulting in a high computational cost and energy consumption level. We propose a new…
Achieving faster execution with shorter compilation time can enable further diversity and innovation in neural networks. However, the current paradigm of executing neural networks either relies on hand-optimized libraries, traditional compilation heuristics, or very recently, simulated annealing and genetic algorithms.…
Improves inference-time alignment for diffusion models without updating weights.
problem Aligning diffusion models without updating weights for high-reward outputs.
method Trust-Region Iterative Twisted Sequential Monte Carlo (TRI-TSMC) for variance reduction and efficiency.
result Improves primary alignment objectives on text generation tasks.
LoRA-MCL improves language models by generating diverse sentence continuations.
problem Language models struggle with generating diverse, plausible sentence continuations.
method Low-Rank Adaptation combined with Multiple Choice Learning (MCL) to handle ambiguity.
result LoRA-MCL generates high-diversity and relevant outputs in various tasks.
PoWER-BERT speeds up BERT inference by eliminating redundant word-vectors.
problem Improving BERT inference speed without sacrificing accuracy.
method Eliminating redundant word-vectors using a self-attention-based significance measure and learning the number of vectors to eliminate.
result Up to 4.5x reduction in inference time with <1% loss in accuracy on GLUE benchmark.
AECF improves multimodal inference robustness and calibration.
problem Robustness and calibration issues in multimodal systems with missing inputs.
method Adaptive Entropy-Gated Contrastive Fusion (AECF) layer.
result Improves masked-input mAP by +18 pp at a 50% drop rate.
In-Place TTT enhances LLMs with dynamic parameter updates at inference time.
problem Static training limits LLMs from adapting to new information.
method In-Place TTT updates a subset of model parameters (fast weights) at inference time.
result In-Place TTT enables 4B-parameter models to outperform on tasks with up to 128k contexts.
Paper speeds up large foundation models for time series data.
problem Resource-intensive foundation models limit accessibility.
method Dimensionality reduction techniques, including PCA and neural network adapters.
result Up to 10x speedup and 4.5x more datasets fit on a single GPU.
New approach tackles class imbalance in long-tailed datasets using domain adaptation techniques.
problem Class imbalance in long-tailed datasets leading to poor model performance.
method Proposes a meta-learning approach to estimate differences between class-conditioned distributions.
result Validated approach on six benchmark datasets and three loss functions.
Chameleon optimizes neural network compilation for faster execution and shorter time.
problem Faster execution and shorter compilation time for neural networks.
method Adaptive code optimization using reinforcement learning and adaptive sampling.
result 4.45x speed up in optimization time over AutoTVM, 5.6% improvement in inference time.
M-FFF generates data on manifolds with fast sampling.
problem Sampling on arbitrary manifolds is computationally expensive.
method Optimizes a neural network via maximum likelihood on the manifold.
result Consistently matches or outperforms previous methods.
As neural networks become widely deployed in different applications and on different hardware, it has become increasingly important to optimize inference time and model size along with model accuracy. Most current techniques optimize model size, model accuracy and inference time in different stages, resulting in subopt…
While several approaches to face emotion recognition task are proposed in literature, none of them reports on power consumption nor inference time required to run the system in an embedded environment. Without adequate knowledge about these factors it is not clear whether we are actually able to provide accurate face e…
A new method for math reasoning that allows for iterative correction.
problem Standard reasoning models commit to each token and cannot recover from early errors.
method Generative framework with latent thought vectors for iterative self-correction.
result 30 rethinking iterations surpass baselines with 15 times more parameters.
PEAR dynamically reconfigures agent roles to prevent persistent biases in multi-agent debates.
problem Persistent positional biases and sensitivity to role assignments in fixed topologies.
method Dynamic reconfiguration of agent roles and sparse topologies based on evolving agent states.
result Significantly improves average accuracy over debate baselines across multiple reasoning benchmarks.
Neural Networks trained with gradient descent are known to be susceptible to catastrophic forgetting caused by parameter shift during the training process. In the context of Neural Machine Translation (NMT) this results in poor performance on heterogeneous datasets and on sub-tasks like rare phrase translation. On the …
TTT improves model adaptation to test data, especially for nonlinear models.
problem Improving model performance in adapting to test data, especially for nonlinear models.
method Combining Test-time Training (TTT) with In-context Learning (ICL) for nonlinear models.
result TTT enables models to adapt to both feature vector and link function shifts, improving performance.
New method improves grouped convolutions on edge devices.
problem Efficiently implementing grouped convolutions on edge devices.
method Grouped Spatial Pack Convolutions (GSPC) in TVM.
result GSPC outperforms existing implementations by 3.4x, 8x, and 4x on average.
We introduce a novel view to understand how dropout works as an inexplicit ensemble learning method, which doesn't point out how many and which nodes to learn a certain feature. We propose a new training method named internal node bagging, it explicitly forces a group of nodes to learn a certain feature in training tim…
Paper organizes sampling methods for generative modeling.
problem Challenges in sampling with diffusion models.
method Reviews and organizes existing sampling methods.
result Reveals links between methods to overcome challenges.
MFMs enable efficient reward alignment for generative models.
problem Computational bottleneck in controlling generative models.
method Meta Flow Maps (MFMs) extend consistency models and flow maps to stochastic regime for efficient value function estimation.
result MFMs enable inference-time steering and unbiased, off-policy fine-tuning to general rewards efficiently.
Efficiently infers time-varying sparse MRFs with strong statistical guarantees.
problem Inference of time-varying sparse MRFs with strong statistical guarantees.
method Constrained optimization with exact ℓ0 regularization, near-linear time and memory complexity. result Sharp statistical guarantees for sparsely-changing Gaussian MRFs with as few as one sample per time.
Paper tackles continual learning with Transformers, maintaining performance and speed.
problem Catastrophic forgetting in continual learning with limited resources.
method Incremental training with pre-trained Transformers and Adapters.
result Maintains good predictive performance and inference speed without retraining.