arXiv subjects

Da Huang

Publications and source records attributed to Da Huang.

At least 19 recordsLinked to original sources

arXiv2026

GE-Act 2.0: Pretraining and Scaling a World-Action Model for Robotic Manipulation

World-action models (WAM) predict future states to guide robot actions, enabling learning from both action-free video and action-labeled interaction. Most inherit pretrained video generators, leaving WAM pretraining and scaling underexplored. We introduce Genie Envisioner Act 2.0 (GE-Act 2.0), a world-action model whose trainable generative and action components are all initialized from scratch on manipulation data. It combines a control-oriented autoencoder (CoAE), a single-step visual planner (SVP), and an inverse dynamics model (IDM). CoAE retains action- and instruction-relevant information under aggressive compression, while SVP produces a complete future state in one differentiable pass, so visual planning and inverse dynamics can be pretrained separately on complementary data. The components are then jointly trained with knowledge-aligned selective optimization (KASO), which reduces mismatched supervision by selecting only predicted futures judged behaviorally compatible with the recorded action. We evaluate pretrained checkpoints directly, without per-task fine-tuning, on 100 tasks across 20 manipulation skill groups with held-out scenes, backgrounds, lighting, and object instances. Scaling co-training data from 300 to 30,000 hours raises success from 17.1% to 44.1% on G1-OP and from 13.4% to 31.1% on G2-90D; despite comprising less than 2% of the co-training data, G2-90D improves by 17.7 points, suggesting cross-embodiment transfer. Gains span 19/20 and 18/20 skill groups, and skill-specific coverage strongly correlates with zero-shot out-of-distribution (OOD) success (Pearson r=0.80; Spearman rho=0.85). Under the same protocol, the model grounds object, color, shape, and position references in at least 90% of trials and follows explicit instructions even when they conflict with an already-committed behavior or a conventional scene association.

cs.RO↗

arXiv2026

Cosmological Constraints on Temperature-Dependent Interaction between Dark Matter and Neutrinos

We study the influence of the temperature-dependent interaction between dark matter (DM) and neutrinos on the measurement of cosmological parameters. We pay attention to the neutrino mass effects, so that the derivation of Boltzmann equations needs to specify the concrete form of interaction. We work in a model in which the DM-neutrino scatterings are induced by a dimension-six operator, and present the details for deriving the full Boltzmann hierarchy for DM and neutrinos, including a novel method to obtain the fluid approximation for modes entering the horizon. It is shown that our interaction can induce the dark acoustic oscillation in the DM-neutrino fluid, leaving distinct signatures on the CMB and matter power spectra. By using the latest CMB and BAO datasets from Planck, DESI and ACT, the constraint on today's DM-neutrino interaction parameter for the normal neutrino mass ordering reaches $u^0_{\chi-\nu} \lesssim {\cal O}(10^{-13})$, nearly nine orders stronger than that for temperature-independent case in the literature. This can be understood by noting that the scattering cross section increases nearly quadratically with cosmological temperature in the early universe, leading to enhanced effects. We have investigated alternative scenarios with different neutrino mass assumptions. In particular, models with degenerate neutrino masses give rise to weaker constraint of $u^0_{\chi-\nu} \lesssim {\cal O}(10^{-11})$, showing the importance to incorporate the realistic neutrino mass ordering in the fits. Finally, when employing the logarithmic flat prior for $u^0_{\chi-\nu}$, we have shown hints to a nonzero interaction at $95\%$ CL by combining Planck, DESI and ACT data.

astro-ph.CO↗

arXiv2026

Genie Sim 3.0 : A High-Fidelity Comprehensive Simulation Platform for Humanoid Robot

The development of robust and generalizable robot learning models is critically contingent upon the availability of large-scale, diverse training data and reliable evaluation benchmarks. Collecting data in the physical world poses prohibitive costs and scalability challenges, and prevailing simulation benchmarks frequently suffer from fragmentation, narrow scope, or insufficient fidelity to enable effective sim-to-real transfer. To address these challenges, we introduce Genie Sim 3.0, a unified simulation platform for robotic manipulation. We present Genie Sim Generator, a large language model (LLM)-powered tool that constructs high-fidelity scenes from natural language instructions. Its principal strength resides in rapid and multi-dimensional generalization, facilitating the synthesis of diverse environments to support scalable data collection and robust policy evaluation. We introduce the first benchmark that pioneers the application of LLM for automated evaluation. It leverages LLM to mass-generate evaluation scenarios and employs Vision-Language Model (VLM) to establish an automated assessment pipeline. We also release an open-source dataset comprising more than 10,000 hours of synthetic data across over 200 tasks. Through systematic experimentation, we validate the robust zero-shot sim-to-real transfer capability of our open-source dataset, demonstrating that synthetic data can server as an effective substitute for real-world data under controlled conditions for scalable policy training. For code and dataset details, please refer to: https://github.com/AgibotTech/genie_sim.

cs.RO↗

arXiv2025

Schr\"odinger's Navigator: Imagining an Ensemble of Futures for Zero-Shot Object Navigation

Zero-shot object navigation (ZSON) requires robots to find target objects in unseen environments without task-specific fine-tuning or pre-built maps, a key capability for general-purpose service robots. Yet methods that perform well in simulation often degrade in cluttered real-world scenes with severe occlusion and latent hazards, where large unseen regions make single-scene inference brittle and unsafe. We propose Schr\"odinger's Navigator, a belief-aware framework that reasons at inference time over multiple trajectory-conditioned imagined 3D futures. Given candidate paths, a trajectory-conditioned 3D world model predicts hypothetical observations and maintains a superposition of plausible scene realizations rather than committing to one map. An adaptive occluder-aware sampler directs imagination to uncertainty-critical regions, while a Future-Aware Value Map (FAVM) aggregates imagined futures for robust, proactive action selection. Experiments in simulation and on a physical Go2 quadruped show that Schr\"odinger's Navigator outperforms strong ZSON baselines, improving hidden-target discovery and risk-aware waypoint selection in occlusion-heavy navigation scenarios. These results highlight imagined 3D futures as a scalable and generalizable strategy for zero-shot navigation in uncertain real-world environments.

cs.RO↗

arXiv2025

TP-MDDN: Task-Preferenced Multi-Demand-Driven Navigation with Autonomous Decision-Making

In daily life, people often move through spaces to find objects that meet their needs, posing a key challenge in embodied AI. Traditional Demand-Driven Navigation (DDN) handles one need at a time but does not reflect the complexity of real-world tasks involving multiple needs and personal choices. To bridge this gap, we introduce Task-Preferenced Multi-Demand-Driven Navigation (TP-MDDN), a new benchmark for long-horizon navigation involving multiple sub-demands with explicit task preferences. To solve TP-MDDN, we propose AWMSystem, an autonomous decision-making system composed of three key modules: BreakLLM (instruction decomposition), LocateLLM (goal selection), and StatusMLLM (task monitoring). For spatial memory, we design MASMap, which combines 3D point cloud accumulation with 2D semantic mapping for accurate and efficient environmental understanding. Our Dual-Tempo action generation framework integrates zero-shot planning with policy-based fine control, and is further supported by an Adaptive Error Corrector that handles failure cases in real time. Experiments demonstrate that our approach outperforms state-of-the-art baselines in both perception accuracy and navigation robustness.

cs.RO↗

arXiv2025

TrajBooster: Boosting Humanoid Whole-Body Manipulation via Trajectory-Centric Learning

Recent Vision-Language-Action models show potential to generalize across embodiments but struggle to quickly align with a new robot's action space when high-quality demonstrations are scarce, especially for bipedal humanoids. We present TrajBooster, a cross-embodiment framework that leverages abundant wheeled-humanoid data to boost bipedal VLA. Our key idea is to use end-effector trajectories as a morphology-agnostic interface. TrajBooster (i) extracts 6D dual-arm end-effector trajectories from real-world wheeled humanoids, (ii) retargets them in simulation to Unitree G1 with a whole-body controller trained via a heuristic-enhanced harmonized online DAgger to lift low-dimensional trajectory references into feasible high-dimensional whole-body actions, and (iii) forms heterogeneous triplets that couple source vision/language with target humanoid-compatible actions to post-pre-train a VLA, followed by only 10 minutes of teleoperation data collection on the target humanoid domain. Deployed on Unitree G1, our policy achieves beyond-tabletop household tasks, enabling squatting, cross-height manipulation, and coordinated whole-body motion with markedly improved robustness and generalization. Results show that TrajBooster allows existing wheeled-humanoid data to efficiently strengthen bipedal humanoid VLA performance, reducing reliance on costly same-embodiment data while enhancing action space understanding and zero-shot skill transfer capabilities. For more details, For more details, please refer to our \href{https://jiachengliu3.github.io/TrajBooster/}.

cs.RO↗

arXiv2025

Gravitational Wave Generation and Detection in Gravitational Quantum Field Theory

We investigate the production and detection of gravitational waves (GWs) within the framework of Gravitational Quantum Field Theory (GQFT). In this theory, GWs exhibit five propagating modes: one scalar, two vector, and two tensor modes. Unlike General Relativity, the gravitational field equations in GQFT involve both symmetric and antisymmetric tensors, governed by their respective energy-momentum tensors, both of which can act as sources for GW radiation. By solving the linearized gravitational equations, we derive general analytic expressions for the different GW degrees of freedom. Our analysis reveals that the symmetric energy-momentum tensor generates scalar and tensor GWs through the trace and traceless parts of the quadrupole moment, respectively. In contrast, the antisymmetric stress tensor induces scalar and vector GWs with enhanced coupling strengths. We examine two illustrative examples: a black hole binary with a slightly elliptical orbit, which produces scalar GWs, and a neutron star binary where one component has a net spin aligned with its velocity, leading to vector GW emission. Finally, we study the detectability of these GW polarizations by analyzing their signatures in GW detectors. Our findings indicate that current observatories can detect both scalar and tensor modes, while a newly designed detector would be required to probe vector GWs.

gr-qc↗

arXiv2024

Gravitational Wave Birefringence in Symmetron Cosmology

The symmetron is a light scalar which provides a screening mechanism so as to evade the strong constraints from local gravity tests. In order to achieve this goal, a $Z_2$ symmetry is imposed on the symmetron model. In this paper, we introduce a new symmetron Chern-Simons-like gravitational interaction which is $Z_2$ invariant but breaks the parity symmetry explicitly. As a result, it is found that this coupling can generate gravitational wave (GW) amplitude birefringence when GWs propagate over the symmetron backgrounds. Due to the matter density difference, the symmetron profile changes significantly when entering the galaxy, so that we need to discuss the extra-galactic and galactic situations separately. On the one hand, the cosmological symmetron field follows the adiabatic solution, which induces a parity-violating GW amplitude correction with its exponent proportional to the GW frequency and the traveling distance. On the other hand, the symmetron takes the screening solution within the Milky Way, and the generated GW birefringence is only a function of the GW frequency. By further comparing these two contributions, we find that the extra-galactic symmetron field produces the dominant birefringence effects. Finally, with the latest GW data from LIGO-Virgo-Kagra, we place a reasonable constraint on the parity-violating coupling parameter in this symmetron model.

gr-qc↗

arXiv2024

Gravitational Wave Birefringence from Fuzzy Dark Matter

Gravitational wave (GW) birefringence is a remarkable phenomenon that can be used to test the parity violation in gravity. By coupling the fuzzy dark matter (FDM) scalar to the gravitational Chern-Simons term, we explore the GW birefringence effects in the FDM background. In particular, in light of the highly oscillating granular FDM structure at the galactic scale, we are led to investigating the GW propagation in the Chern-Simons gravity over the general nontrivial scalar profile, which is a natural extension of previous studies on the homogeneous and isotropic configurations. As a result, it is found that GWs of both circularly polarized modes propagate in the straight line with the speed of light, and does not show any velocity birefringence. However, when considering the imaginary part of the dispersion relation, GWs exhibit the amplitude birefringence in which one circular polarization is enhanced while the other suppressed. Due to its local nature, the FDM-induced amplitude birefringence factor only depends on the GW frequency without any reliance on the GW propagating distance, which can be used to distinguish this signal from other birefringece mechanisms. More importantly, the birefringence shows a periodic time modulation with the period directly reflecting the FDM scalar mass, which is another smoking gun for testing this model. Finally, we also study the extra-galactic FDM contribution to the GW birefringence, which is shown to be suppressed by the cosmological DM density and thus subdominant compared with the galactic counterpart.

gr-qc↗

arXiv2024

Long-form factuality in large language models

Large language models (LLMs) often generate content that contains factual errors when responding to fact-seeking prompts on open-ended topics. To benchmark a model's long-form factuality in open domains, we first use GPT-4 to generate LongFact, a prompt set comprising thousands of questions spanning 38 topics. We then propose that LLM agents can be used as automated evaluators for long-form factuality through a method which we call Search-Augmented Factuality Evaluator (SAFE). SAFE utilizes an LLM to break down a long-form response into a set of individual facts and to evaluate the accuracy of each fact using a multi-step reasoning process comprising sending search queries to Google Search and determining whether a fact is supported by the search results. Furthermore, we propose extending F1 score as an aggregated metric for long-form factuality. To do so, we balance the percentage of supported facts in a response (precision) with the percentage of provided facts relative to a hyperparameter representing a user's preferred response length (recall). Empirically, we demonstrate that LLM agents can outperform crowdsourced human annotators - on a set of ~16k individual facts, SAFE agrees with crowdsourced human annotators 72% of the time, and on a random subset of 100 disagreement cases, SAFE wins 76% of the time. At the same time, SAFE is more than 20 times cheaper than human annotators. We also benchmark thirteen language models on LongFact across four model families (Gemini, GPT, Claude, and PaLM-2), finding that larger language models generally achieve better long-form factuality. LongFact, SAFE, and all experimental code are available at https://github.com/google-deepmind/long-form-factuality.

cs.CL↗

arXiv2024

Linear dynamics and classical tests of the gravitational quantum field theory

We explore the new physics phenomena of gravidynamics governed by the inhomogeneous spin gauge symmetry based on the gravitational quantum field theory. Such a gravidynamics enables us to derive the generalized Einstein equation and an equation beyond it. To simplify the analyses, we linearize the dynamic equations of gravitational interaction by keeping terms up to the leading order in the dual gravigauge field. We then apply the linearized dynamic equations into two particular gravitational phenomena. First, we consider the linearized equations in the absence of source fields, which is shown to have five physical propagating polarizations as gravitational waves, i.e., two tensor modes, two vector modes, and one scalar, instead of two tensor polarizations in the general relativity. Second, we examine the Newtonian limit in which the gravitational fields and the matter source distribution are weak and static. By deriving the associated Poisson equation, we obtain the exact relation of the fundamental interaction coupling in the gravidynamics with the experimentally measured Newtonian constant. We also make use of nonrelativistic objects and relativistic photons to probe the Newtonian field configurations. In particular, the experiments from the gravitational deflection of light rays and the Shapiro time delay can place stringent constraints on the linearized gravidynamics in the gravitational quantum field theory.

gr-qc↗

arXiv2023

Effects of stimulated emission and superradiant growth of non-spherical axion cluster

We explore the stimulated emission of photons in non-spherical axion clusters with or without the axion source from the superradiance of a rotating black hole (BH). In particular, we focus on the cluster with the initial axion distribution in the $(l,m)=(1,1)$ mode which mimics the shape of an axion cloud induced by the BH superradiance. After establishing the hierarchy of Boltzmann equations governing a general non-spherical axion-photon system, we examine the evolution of photon and axion distributions in the cluster and possible stimulated emission signals. In the case without the axion source, the resultant signal would be a single photon pulse. As for the system with the BH superradiance as the axion source, multiple pulses are predicted. We also show that, for the latter case, the combined effects of stimulated emissions and the axion production from the BH superradiance could reach a balance where the axion cluster becomes uniformly and spherically distributed. Due to the energy and temporal characteristics of the obtained pulses, we demonstrate that the stimulated emissions from the axion cluster with axions sourced by the BH superradiance provide a candidate explanation to the observed fast radio bursts.

hep-ph↗

arXiv2023

Simple synthetic data reduces sycophancy in large language models

Sycophancy is an undesirable behavior where models tailor their responses to follow a human user's view even when that view is not objectively correct (e.g., adapting liberal views once a user reveals that they are liberal). In this paper, we study the prevalence of sycophancy in language models and propose a simple synthetic-data intervention to reduce this behavior. First, on a set of three sycophancy tasks (Perez et al., 2022) where models are asked for an opinion on statements with no correct answers (e.g., politics), we observe that both model scaling and instruction tuning significantly increase sycophancy for PaLM models up to 540B parameters. Second, we extend sycophancy evaluations to simple addition statements that are objectively incorrect, finding that despite knowing that these statements are wrong, language models will still agree with them if the user does as well. To reduce sycophancy, we present a straightforward synthetic-data intervention that takes public NLP tasks and encourages models to be robust to user opinions on these tasks. Adding these data in a lightweight finetuning step can significantly reduce sycophantic behavior on held-out prompts. Code for generating synthetic data for intervention can be found at https://github.com/google/sycophancy-intervention.

cs.CL↗

arXiv2023

$W$-boson Mass Anomaly from High-Dimensional Scalar Multiplets

In light of the recently discovered $W$-boson mass anomaly by the CDF Collaboration, we discuss two distinct mechanisms that could possibly explain this anomaly through the introduction of high-dimensional $SU(2)_L$ scalar multiplets. The first mechanism is the tree-level $W$-boson mass correction induced by the vacuum expectation values of one or more $SU(2)_L$ scalar multiplets with odd dimensions of $n\geq 3$ and zero hypercharge of $Y=0$ in order to avoid the strong constraint from measurements of the $Z$-boson mass. The second mechanism is to consider the one-loop level $W$-boson mass correction from a complex multiplet. In particular, we focus on the case with an additional scalar octuplet with $Y=7/2$. As a result, we find that both mechanisms can explain the $W$-boson mass anomaly without violating any other theoretical or experimental constraints.

hep-ph↗

arXiv2023

Deep Learning Methods for Device Identification Using Symbols Trace Plot

Devices authentication is one crucial aspect of any communication system. Recently, the physical layer approach radio frequency (RF) fingerprinting has gained increased interest as it provides an extra layer of security without requiring additional components. In this work, we propose an RF fingerprinting based transmitter authentication approach density trace plot (DTP) to exploit device-identifiable fingerprints. By considering IQ imbalance solely as the feature source, DTP can efficiently extract device-identifiable fingerprints from symbol transition trajectories and density center drifts. In total, three DTP modalities based on constellation, eye and phase traces are respectively generated and tested against three deep learning classifiers: the 2D-CNN, 2D-CNN+biLSTM and 3D-CNN. The feasibility of these DTP and classifier pairs is verified using a practical dataset collected from the ADALM-PLUTO software-defined radios (SDRs).

eess.SP↗

arXiv2023

Identification Methods for Ordinal Potential Differential Games

This paper introduces two new identification methods for linear quadratic (LQ) ordinal potential differential games (OPDGs). Potential games are notable for their benefits, such as the computability and guaranteed existence of Nash Equilibria. While previous research has analyzed ordinal potential static games, their applicability to various engineering applications remains limited. Despite the earlier introduction of OPDGs, a systematic method for identifying a potential game for a given LQ differential game has not yet been developed. To address this gap, we propose two identification methods to provide the quadratic potential cost function for a given LQ differential game. Both methods are based on linear matrix inequalities (LMIs). The first method aims to minimize the condition number of the potential cost function's parameters, offering a faster and more precise technique compared to earlier solutions. In addition, we present an evaluation of the feasibility of the structural requirements of the system. The second method, with a less rigid formulation, can identify LQ OPDGs in cases where the first method fails. These novel identification methods are verified through simulations, demonstrating their advantages and potential in designing and analyzing cooperative control systems.

math.DS↗

arXiv2023

Brainformers: Trading Simplicity for Efficiency

Transformers are central to recent successes in natural language processing and computer vision. Transformers have a mostly uniform backbone where layers alternate between feed-forward and self-attention in order to build a deep network. Here we investigate this design choice and find that more complex blocks that have different permutations of layer primitives can be more efficient. Using this insight, we develop a complex block, named Brainformer, that consists of a diverse sets of layers such as sparsely gated feed-forward layers, dense feed-forward layers, attention layers, and various forms of layer normalization and activation functions. Brainformer consistently outperforms the state-of-the-art dense and sparse Transformers, in terms of both quality and efficiency. A Brainformer model with 8 billion activated parameters per token demonstrates 2x faster training convergence and 5x faster step time compared to its GLaM counterpart. In downstream task evaluation, Brainformer also demonstrates a 3% higher SuperGLUE score with fine-tuning compared to GLaM with a similar number of activated parameters. Finally, Brainformer largely outperforms a Primer dense model derived with NAS with similar computation per token on fewshot evaluations.

cs.LG↗

arXiv2023

Symbol tuning improves in-context learning in language models

We present symbol tuning - finetuning language models on in-context input-label pairs where natural language labels (e.g., "positive/negative sentiment") are replaced with arbitrary symbols (e.g., "foo/bar"). Symbol tuning leverages the intuition that when a model cannot use instructions or natural language labels to figure out a task, it must instead do so by learning the input-label mappings. We experiment with symbol tuning across Flan-PaLM models up to 540B parameters and observe benefits across various settings. First, symbol tuning boosts performance on unseen in-context learning tasks and is much more robust to underspecified prompts, such as those without instructions or without natural language labels. Second, symbol-tuned models are much stronger at algorithmic reasoning tasks, with up to 18.2% better performance on the List Functions benchmark and up to 15.3% better performance on the Simple Turing Concepts benchmark. Finally, symbol-tuned models show large improvements in following flipped-labels presented in-context, meaning that they are more capable of using in-context information to override prior semantic knowledge.

cs.CL↗