Search arXiv⌕ Search

arXiv subjects

Guannan Qu

Publications and source records attributed to Guannan Qu.

At least 19 recordsLinked to original sources
arXiv2026

Activation-Keyed Momentum: An Anisotropic Momentum Update via the Delta Rule

Most modern optimizers form their momentum as an exponential moving average (EMA) of past gradients, forgetting every direction at one fixed rate. However, the inputs a deep network sees during training can be highly anisotropic, with a few directions queried frequently while most are seen rarely. Preconditioning methods address this anisotropy by wrapping extra processing around this buffer and leave the momentum update itself unchanged. We propose Activation-Keyed Momentum (AK-Momentum), which builds direction-awareness into the momentum update rule. The gradient of a linear layer splits into an input activation that acts as a $\textit{key}$ and an output-side error that acts as a $\textit{value}$. Keying on that activation, AK-Momentum updates the momentum buffer by the canonical delta rule, so each direction is forgotten at a rate set by how often it appears. We prove that it is a valid momentum, that it applies the input-side curvature correction without matrix inversion, and that it clears stale directions faster than EMA under both a fixed and a drifting optimum. It is a drop-in replacement for the momentum buffer of any optimizer, its coefficient transfers across widths under $μ$P, and its extra compute stays between $22.2\%$ and $25.0\%$ of a gated-MLP block's linear cost with no persistent memory. In FineWeb-Edu pretraining, AdamW with AK-Momentum (AK-AdamW) reaches AdamW's validation loss in up to $46.64 \pm 4.52\%$ fewer steps at 67M and $22.14 \pm 0.81\%$ at 370M over three seeds, and the gain persists at 1B on a Chinchilla-optimal budget. A Muon baseline tuned under the same protocol sits above AK-AdamW at both language-model scales, and the gain holds for SGD, ResNet-18, and ViT-Tiny on CIFAR-10. Training-time diagnostics confirm the predicted mechanism, better gradient tracking and healthier input directions. Code is available at https://github.com/euijinh/AK-Momentum.

cs.LG↗
arXiv2026

When Do Discrete Actions Make Good Continuous Controllers?

Though modern AI models like LLMs predict discrete tokens, today's robotic foundation models represent actions as continuous vectors. In this paper, we challenge this common consensus by asking if continuous action parametrizations are superior to their discrete counterparts. We conduct an extensive comparison of single-step regression and multi-step flow variants of both continuous and discrete policies in both from-scratch single-task and multi-task VLA training, in both simulation and on real hardware. In all regimes, we find that discrete action prediction, even the most naive, uniform-bin quantization, can match or even outperform continuous-action policies in task success rate. Moreover, we find that the challenge to discrete action policies is not action precision, but rather factorization error introduced by product-distribution predictions across time and action dimensions. We show that, when paired with the correct interventions (iterative computation through discrete flows, or distribution sharpening through classifier-free guidance), discrete flow policies can turn this weakness into a strength, enabling more responsive action changes during a predicted action chunk. Lastly, we show that attempts to explicitly inject geometric information into discrete policies, e.g. through corruption processes that emulate continuous flow matching, do not yield consistent improvements. Taken together, our findings question existing assumptions that robotic action prediction requires distinct design decisions compared to other AI domains, and motivate the possibility of discrete action prediction as a compelling alternative to today's continuous action parametrizations. Project page: https://iscoyizj.github.io/discrete-action-website/

cs.RO↗
arXiv2026

Ask to Learn: Query-Based Collaborative Reinforcement Learning

Offline reinforcement learning is fundamentally limited by the coverage of the available data, while collecting additional trajectories through online exploration may be costly, unsafe, or infeasible. We study whether this limitation can be overcome without directly interacting with the underlying environment. Motivated by practical settings such as collaborator-mediated simulation, where a collaborator can provide information complementary to the offline data, we introduce query-based collaborative reinforcement learning to study how to acquire this information efficiently. The learner starts with a fixed offline dataset and adaptively queries state-time pairs; the collaborator samples an action according to a fixed, possibly suboptimal policy, and the simulator returns the corresponding reward and next state. For finite-horizon tabular problems, we derive an instance-dependent information-theoretic lower bound on the additional queries required to identify an $ε$-optimal policy and develop a local quadratic approximation. Guided by this characterization, we develop an adaptive query-allocation algorithm whose query complexity nearly matches the local benchmark under suitable conditions. Experiments on multiple synthetic and semi-synthetic environments demonstrate the empirical efficiency of the proposed method.

stat.ML↗
arXiv2026

SWAP: Test-Time Steerable Humanoid Whole-Body Collision Avoidance

Humanoid robots operating in human-centric environments must coordinate the entire body to traverse tight, cluttered spaces while maintaining balance. Existing methods often struggle to jointly consider dynamic stability and collision avoidance from high-dimensional perceptual inputs. Moreover, for humanoids, safe behavior is inherently multimodal: the same obstacle may admit multiple valid traversals, which is difficult to represent with a single deterministic policy. To address these challenges, we propose SWAP (Steerable Whole-body Avoidance Policy), a hierarchical framework that separates collision-aware decision making from stabilization. A high-level, perception-conditioned diffusion policy generates multimodal, collision-aware motion proposals, while a low-level whole-body RL controller tracks these proposals and maintains dynamic balance. To efficiently train the high-level policy, we obtain scalable supervision via hindsight labeling: we roll out the low-level policy under randomized commands, place obstacles only in the space that each executed motion did not sweep, so the motion is collision-free by construction, and re-render the perception the robot would have seen. At deployment, Monte Carlo sampling with simple cost signals steers the diffusion model toward safe, task-aligned behaviors without heavy reward tuning. Experiments in simulation and on a real Unitree G1 humanoid show reliable avoidance with static and dynamic obstacles. Test-time steering further enables customization, improving targeted rewards (e.g., speed, orientation, or posture) by 34% on average while maintaining a high collision avoidance rate. Together, these results support a practical solution to humanoid whole-body collision avoidance with test-time steerability. Project website: https://iscoyizj.github.io/SWAP-steerable-humanoid-oa-website/

cs.RO↗
arXiv2026

HilbertGen-3D:Hilbert--Multifractal Conditioning for Topology-Aware 3D Generation

High-fidelity 3D generative modeling underpins game asset creation, medical imaging, and VR/AR environmental design, where geometric structure directly influences perception and decision-making. Diffusion models represent a powerful paradigm for point cloud generation. However, existing approaches rely heavily on local voxel or patch features and often miss global structural organization, and much less work has studied how global conditioning can be used to improve generation quality. In this paper, we introduce HilbertGen-3D, a novel topology-aware conditional diffusion framework based on Hilbert-Multifractal Conditioning (HMC). HMC computes multi-scale voxel measures, serializes them via a 3D Hilbert curve to expose structured spatial dependencies, and estimates the generalized fractal dimensions from scale-dependent partition functions. We encode these signals as global embeddings and multi-scale tokens, and inject them into a diffusion denoiser through an efficient bottleneck fusion module. Across the Chair, Airplane, and Car categories of ShapeNet, our two conditioning variants reduce the absolute deviation of 1-Nearest Neighbor Accuracy (1-NNA) from 50% by up to 8.90 percentage points and increase coverage by up to 5.88 percentage points relative to TopoDiT-3D. This shows that combining local density cues with global Hilbert-multifractal structure provides a simple and effective recipe for high-fidelity and diverse 3D generation.

cs.CV↗
arXiv2026

Towards Effective Theory of LLMs: A Representation Learning Approach

We propose Representational Effective Theory (RET), a framework for describing large language model computation in terms of learned macrostates rather than microscopic details. RET learns these macrostates from hidden-state trajectories using a BYOL/JEPA-style self-supervised objective, coarse-graining activations into macrovariables that preserve higher-level structure relevant for prediction and interpretation. We evaluate whether these macrovariables are practically relevant for interpretability: RET yields temporally consistent states that reveal ``mental-state'' trajectories of reasoning, capture high-level semantic structure, support early prediction of behavioral outcomes such as sycophancy and alignment faking, while providing causal handles for steering generations toward interpretable computational phases. Together, these results suggest that LLM computation admits useful effective descriptions via RET: high-level, dynamically meaningful variables for interpretation, prediction, and control.

cs.LG↗
arXiv2026

TRAP: Understanding and Mitigating Privacy Memorization in Language Models

Fine-tuning a language model on sensitive records can leave it able to reproduce them. We ask when this memorization arises and how to prevent it without knowing in advance which spans are sensitive. Our starting point is that most memorization scores and attacks share one statistical core: whether the model assigns a token more probability than some reference would. Taking as the reference a model trained on the complementary half of the same corpus gives the Target Reference Advantage (TRA), a per-token signal that separates what a model fit to a particular record from what it learned across records, and is cheap and differentiable. We then study what drives memorization during fine-tuning: it keeps growing well past the validation minimum, is larger on small datasets and at higher learning rates, and higher when the underlying task is harder. Early stopping removes much of it, but because it is chosen by aggregate validation loss it helps least for rare, hard-to-predict spans embedded in otherwise learnable text, which is exactly what sensitive information tends to be. We therefore introduce TRAP, a one-sided penalty on tokenwise TRA that acts only where the target model pulls ahead of its reference. On student essays with annotated personal information and clinical cases with patient identifiers, TRAP brings memorization near the level of an untrained model at little utility cost, where generic regularizers barely move and differential privacy gives up most of what fine-tuning bought.

cs.CL↗
arXiv2026

Learning Fractional-Order Dynamics from a Single Trajectory

Many real-world processes exhibit long-range dependence, where the current state depends on a slowly decaying trace of past states rather than on the most recent state alone. This paper studies system identification for discrete-time fractional-order linear time-invariant systems from a single observed trajectory of length $t$, a setting that captures such non-Markovian dynamics through the Grünwald--Letnikov difference operator. Unlike Markovian systems, fractional-order systems couple estimation across the entire history, making both statistical analysis and practical identification more challenging. We propose \emph{Fractional-Order Ordinary-Least-Squares Grid-Search (FO-GS)}, a simple two-stage estimator that exploits the diagonal structure of the fractional-difference operator to decouple the identification problem row-wise. Under the stability assumption, we establish high-probability, non-asymptotic error bounds for estimating both the fractional order and the system matrix in the heterogeneous setting, with both estimation errors scaling as \(\mathcal{O}(t^{-1/2})\). Through experiments, we show that \emph{FO-GS} outperforms existing baselines in recovering both the fractional order and the underlying system dynamics.

cs.LG↗
arXiv2026

Hybrid Feedback Sampling for Sample-Efficient Model Predictive Control

Thanks to its parallelizability and flexibility, sampling-based Model Predictive Control (MPC) has become widely popular for controlling real-world robotic systems. However, for high-dimensional and open-loop unstable dynamical systems, the required number of samples to improve the control sequence will grow exponentially with the horizon, leading to poor sample efficiency and numerical instability. This paper investigates the instability of shooting methods in sampling-based MPC and shows that the optimal sampling proposal distribution can be realized by sampling with an optimized feedback policy. We refer to this algorithm as Feedback Sampling MPC (FS-MPC). FS-MPC involves a hybrid sampling design which balances local and global search based on the system stability and the available computation budget. Our theoretical analysis shows that our hybrid sampling approach achieves faster convergence than standard MPPI and better optimality than standard feedback sampling. Empirically, in diverse contact-rich control tasks like humanoid loco-manipulation and dexterous manipulation, we show that FS-MPC successfully tackles dynamically unstable tasks where standard sample-based approaches struggle, and strictly outperforms feedback policies alone. Finally, we validate our method on humanoid robot locomotion and manipulation tasks in the real world.

cs.RO↗
arXiv2026

A Theoretical Framework for Parallel Lifelong MAPF Using Group Decentralized Planning

In the Lifelong Multi-Agent Path Finding (L-MAPF) problem, agents must repeatedly move from one destination to another while avoiding obstacles and inter-agent collisions. Widely regarded as one of the highest-performing solutions to this problem is the Rolling-Horizon Collision Resolution (RHCR) framework. However, commensurate with its quality solutions, it incurs a computational cost that limits its applicability to even modest agent counts. In this paper, leveraging theoretical methods from the Locally Interdependent Multi-Agent MDP literature, we first theoretically prove the near-optimality of RHCR in a discounted MDP formulation of the L-MAPF problem. Then, we leverage these results to naturally motivate an extended framework called Group Decentralized RHCR (GD-RHCR) which incorporates a group decentralized structure that partitions agents based on a transitive communication scheme and plans for each partition of agents in parallel. We show that both RHCR and GD-RHCR achieve similar exponentially close to optimal guarantees, establishing a theoretical duality between the time based restrictions performed by vanilla RHCR and the additional space based partitioning performed by GD-RHCR. Lastly, we show that across varying maps, GD-RHCR is able to attain high throughput that scales into higher agent counts while maintaining a significantly lower per plan cost.

cs.MA↗
arXiv2026

Comparative Field Deployment of Reinforcement Learning and Model Predictive Control for Residential HVAC

Model Predictive Control (MPC) has demonstrated significant performance improvements over today's control methods for residential Heating, Ventilation, and Air Conditioning (HVAC), but deploying MPC often requires substantial engineering effort. Reinforcement Learning (RL) may offer comparable performance with easier deployment, but its practical application for residential HVAC remains largely undemonstrated, leaving open questions related to occupant comfort and data requirements. To investigate these issues, we deployed one MPC variant and one model-based RL variant for one month each in an occupied house in a cold climate. The controllers adjusted an air-to-air heat pump's thermostat temperature setpoint based on measurements of the indoor temperature and the electric power used for heating. Relative to constant-setpoint operation, MPC saved 18.1\% (95\% confidence interval: 4.4 to 30.9\%) of weather-normalized heat pump energy and RL saved 20.9\% (2.6 to 38.3\%). MPC maintained acceptable occupant comfort. RL kept the house cooler, particularly during an initial adaptation phase, leading to three reports of occupant discomfort. The two algorithms had similar data requirements. We estimate that for a fresh deployment in another house, RL would take about one-third less engineering effort than MPC. While RL reduces deployment effort, it faces difficulties related to safe controller initialization and to mismatches between the modeled and true state and action spaces.

eess.SY↗
arXiv2026

Transformer-Based Multi-Agent Reinforcement Learning for Networked Systems with Long-Range Interactions

Multi-agent reinforcement learning (MARL) has shown promise for large-scale network control, yet existing methods face two major limitations. First, they typically rely on an exponential decay property of agent interactions on far-away nodes, which can be exploited to develop more efficient and tractable MARL algorithms. When this exponential decay property does not hold, these algorithms do not account for long-range interactions such as epidemic outbreaks or cascading power failures. Second, existing approaches lack network generalizability, or the ability to generalize to networks of different topological structure and scale than those seen during training. In this work, we first present a mean-field stability analysis and empirical study investigating the conditions for long-range network interactions. These results motivate our primary contribution: STACCA (Shared Transformer Actor-Critic with Counterfactual Advantage), a transformer-based MARL framework that addresses both long-range interactions and network generalizability. STACCA employs a centralized Graph Transformer Critic to model long-range dependencies and provide system-level feedback, while its shared Graph Transformer Actor learns a generalizable policy capable of adapting across diverse network topologies. To improve credit assignment during training, STACCA integrates a novel counterfactual advantage estimator that is compatible with state-value critic estimates. We evaluate STACCA on epidemic containment and rumor-spreading network control tasks, demonstrating improved performance and network generalizability. These results highlight the potential of transformer-based MARL architectures to achieve generalizable control in large-scale networked systems.

cs.LG↗
arXiv2026

Global Convergence of Sampling-Based Nonconvex Optimization through Diffusion-Style Smoothing

Sampling-based optimization (SBO), like cross-entropy method and evolutionary algorithms, has achieved many successes in solving non-convex problems without gradients, yet its convergence is poorly understood. In this paper, we establish a non-asymptotic convergence analysis for SBO through the lens of smoothing. Specifically, we recast SBO as gradient descent on a smoothed objective, mirroring noise-conditioned score ascent in diffusion models. Our first contribution is a landscape analysis of the smoothed objective, demonstrating how smoothing helps escape local minima and uncovering a fundamental coverage-optimality trade-off: smoothing renders the landscape more benign by enlarging the locally convex region around the global minimizer, but at the cost of introducing an optimality gap. Building on this insight, we establish non-asymptotic convergence guarantees for SBO algorithms to a neighborhood of the global minimizer. Furthermore, we propose an annealed SBO algorithm, Diffusion-Inspired Dual-Annealing (DIDA), which is provably convergent to the global optimum. We conduct extensive numerical experiments to verify our landscape results and also demonstrate the compelling performance of DIDA compared to other gradient-free optimization methods. Lastly, we discuss implications of our results for diffusion models.

cs.LG↗
arXiv2026

BLOCK-EM: Preventing Emergent Misalignment via Latent Blocking

Emergent misalignment can arise when a language model is fine-tuned on a narrowly scoped supervised objective: the model learns the target behavior, yet also develops undesirable out-of-domain behaviors. We investigate a mechanistic approach to preventing emergent misalignment by identifying a small set of internal features that reliably control the misaligned behavior and then discouraging the model from strengthening these features during fine-tuning. Across six fine-tuning domains, blocking (i.e., constraining) a fixed set of features achieves up to 95\% relative reduction in emergent misalignment with no degradation in model quality or target-task performance. We strengthen validity with disjoint selection/evaluation splits, multiple independent judges, multiple random seeds for key settings, quality metrics, and extensive ablations demonstrating that the reduction in misalignment is specific to the identified mechanism. We also characterize a limiting regime in which misalignment re-emerges under prolonged fine-tuning, present evidence consistent with rerouting through alternative features or layers, and evaluate modifications that partially restore the misalignment-blocking effect. Overall, our results show that targeted training-time constraints on internal mechanisms can mitigate emergent misalignment without degrading target-task performance.

cs.LG↗
arXiv2026

Emergent and Subliminal Misalignment Through the Lens of Data-Mediated Transfer

Fine-tuning LLMs on narrow harmful datasets can induce Emergent Misalignment (EM), where models exhibit misaligned behavior far beyond the fine-tuning distribution. We argue that emergent misalignment can be better understood as a data-mediated transfer phenomenon: harmful fine-tuning examples do not induce uniform behavioral spillover, but interact with the structural properties of the dataset and the difficulty of the tasks relative to the model. Across our experiments, we find that misalignment appears more readily when fine-tuning and evaluation prompts share similar underlying functional structure, when prompts leave more room for coherent harmful completions, and when the target behavior has been more reliably learned by the model. The training pipeline itself also matters: pretraining composition shapes later misalignment. We further study Subliminal Learning (SL), where misalignment is transmitted by fine-tuning on seemingly benign data generated by a harmful teacher. Moving beyond the standard SFT setting, we for the first time compare this transfer under off-policy and on-policy distillation as well, allowing us to separate the roles of the teacher guidance and the training data distribution in transmitting misalignment. Together, these results argue for a data-centric view: Emergent/subliminal misalignment should not be treated as a simple consequence of isolated harmful fine-tuning examples, but as the result of interactions between fine-tuning data structure, pretraining distributions, and training channels.

cs.LG↗
arXiv2026

Revisiting Policy Gradients for Restricted Policy Classes: Escaping Myopic Local Optima with $k$-step Policy Gradients

This work revisits standard policy gradient methods used on restricted policy classes, which are known to get stuck in suboptimal critical points. We identify an important cause for this phenomenon to be that the policy gradient is itself fundamentally myopic, i.e. it only improves the policy based on the one-step $Q$-function. In this work, we propose a generalized $k$-step policy gradient method that couples the randomness within a $k$-step time window and can escape the myopic local optima in MDPs with restricted policy classes. We show this new method is theoretically guaranteed to converge to a solution that is exponentially close in performance to the optimal deterministic policy with respect to $k$. Further, we show projected gradient descent and mirror descent with this $k$-step policy gradient can achieve this exponential guarantee in $O(\frac{1}{T})$ iterations, despite only assuming smoothness and differentiability of the value function. This will provide near optimal solutions to previously elusive applications like state aggregation and partially observable cooperative multi-agent settings. Moreover, our bounds avoid the ubiquitous distribution mismatch factors $||d_μ^{π^*} / d_μ^π||_\infty$ and $||d_μ^{π^*} / μ||_\infty$ enabling the $k$-step policy gradient method to escape suboptimal critical points that emerge from poor exploration in fully observable settings.

cs.LG↗
arXiv2026

Multi-Period Sparse Optimization for Proactive Grid Blackout Diagnosis

Existing or planned power grids need to evaluate survivability under extreme events, like a number of peak load overloading conditions, which could possibly cause system collapses (i.e. blackouts). For realistic extreme events that are correlated or share similar patterns, it is reasonable to expect that the dominant vulnerability or failure sources behind them share the same locations but with different severity. Early warning diagnosis that proactively identifies the key vulnerabilities responsible for a number of system collapses of interest can significantly enhance resilience. This paper proposes a multi-period sparse optimization method, enabling the discovery of persistent failure sources across a sequence of collapsed systems with increasing system stress, such as rising demand or worsening contingencies. This work defines persistency and efficiently integrates persistency constraints to capture the ``hidden'' evolving vulnerabilities. Circuit-theory based power flow formulations and circuit-inspired optimization heuristics are used to facilitate the scalability of the method. Experiments on benchmark systems show that the method reliably tracks persistent vulnerability locations under increasing load stress, and solves with scalability to large systems (on average taking around 200 s per scenario on 2000+ bus systems).

eess.SY↗
arXiv2026

Whole-Body Model-Predictive Control of Legged Robots with MuJoCo

We demonstrate the surprising real-world effectiveness of a very simple approach to whole-body model-predictive control (MPC) of quadruped and humanoid robots: the iterative LQR (iLQR) algorithm with MuJoCo dynamics and finite-difference approximated derivatives. Building upon the previous success of model-based behavior synthesis and control of locomotion and manipulation tasks with MuJoCo in simulation, we show that these policies can easily generalize to the real world with few sim-to-real considerations. Our baseline method achieves real-time whole-body MPC on a variety of hardware experiments, including dynamic quadruped locomotion, quadruped walking on two legs, and full-sized humanoid bipedal locomotion. We hope this easy-to-reproduce hardware baseline lowers the barrier to entry for real-world whole-body MPC research and contributes to accelerating research velocity in the community. Our code and experiment videos will be available online at:https://johnzhang3.github.io/mujoco_ilqr

cs.RO↗