arXiv subjects

Haitao Cui

Publications and source records attributed to Haitao Cui.

4 recordsLinked to original sources

arXiv2026

DeepInsight: A Unified Evaluation Infrastructure Across the Physical AI Stack

Evaluating a Physical AI stack spans operators that differ by more than three orders of magnitude -- from a single foundation-model decoding step to thousands of physics ticks of whole-body control -- varying orthogonally in modality, reward semantics, and resource profile. No existing framework spans this range, so the stack is evaluated today by stitching together separate harnesses that share neither runtime nor scoring, preserving each segment's local validity but losing the shared identity needed to diagnose cross-layer regressions. We present DeepInsight, an evaluation infrastructure that serves this full spectrum on a single runtime. Rather than homogenize the regimes, it preserves their heterogeneity behind three narrow abstractions -- task, resource, and result -- each realized as one invariant shared by every subsystem: one episode driver, one resource-handle protocol implemented by every expensive backend (LLM inference and sandboxed runtimes alike), and one trace identity scheme under which every event is written. Deployed in production across all three layers of an embodied humanoid stack, this single set of invariants onboards new benchmarks largely by configuration. Where mature peer orchestrators exist -- at the foundation-model end -- it reproduces published references and peer-framework readings within their own spread, runs the same suites faster on a single node, and scales near-linearly across nodes. Its distinctive return is diagnostic: because every layer writes into one shared trace, a regression that begins in one layer and surfaces in another stays localizable on that trace -- a cross-layer payoff no federation of per-segment harnesses can reproduce.

cs.AI↗

arXiv2017

The Strong Cell-based Hydrogen Peroxide Generation Triggered by Cold Atmospheric Plasma

Hydrogen peroxide (H2O2) is an important signaling molecule in cancer cells. However, the significant secretion of H2O2 by cancer cells have been rarely observed. Cold atmospheric plasma (CAP) is a near room temperature ionized gas composed of neutral particles, charged particles, reactive species, and electrons. Here, we first demonstrated that breast cancer cells and pancreatic adenocarcinoma cells generated micromolar level H2O2 during just 1 min of direct CAP treatment on these cells. The cell-based H2O2 generation is affected by the medium volume, the cell confluence, as well as the discharge voltage. The application of cold atmospheric plasma (CAP) in the cancer treatment has been intensively investigated over the past decade. Several cellular responses to the CAP treatment have been observed including the consumption of the CAP-originated reactive species, the rise of intracellular reactive oxygen species, the damage on DNA and mitochondria, as well as the activation of apoptotic events. This is a new previously unknown cellular response to CAP, which provides a new prospective to understand the interaction between CAP and cells.

q-bio.CB↗

arXiv2016

Effect of Wigner energy on the symmetry energy coefficient in nuclei

The nuclear symmetry energy coefficient (including the coefficient $a_{\rm sym}^{(4)}$ of $I^{4}$ term) of finite nuclei is extracted by using the differences of available experimental binding energies of isobaric nuclei. It is found that the extracted symmetry energy coefficient $a^{*}_{\rm sym}(A,I)$ decreases with increasing of isospin asymmetry $I$, which is mainly caused by Wigner correction, since $e^{*}_{\rm sym}$ is the summation of the traditional symmetry energy $e_{\rm sym}$ and the Wigner energy $e_{\rm W}$. We obtain the optimal values $J=30.25\pm0.10$ MeV, $a_{\rm ss}=56.18\pm1.25$ MeV, $a_{\rm sym}^{(4)}=8.33\pm1.21$ MeV and the Wigner parameter $x=2.38\pm0.12$ through the polynomial fit to 2240 measured binding energies for nuclei with $20 \leq A \leq 261$ with an rms deviation of 23.42 keV. We also find that the volume symmetry coefficient $J\simeq 30$ MeV is insensitive to the value $x$, whereas the surface symmetry coefficient $a_{\rm ss}$ and the coefficient $a_{\rm sym}^{(4)}$ are very sensitive to the value of $x$ in the range $1\leq x\leq 4$. The contribution of $a_{\rm sym}^{(4)}$ term increases rapidly with increasing of isospin asymmetry $I$. For very neutron-rich nuclei, the contribution of $a_{\rm sym}^{(4)}$ term will play an important role.

nucl-th↗

arXiv2014

Extraction of the symmetry energy coefficients from the masses differences of isobaric nuclei

The nuclear symmetry energy coefficients of finite nuclei are extracted by using the differences between the masses of isobaric nuclei. Based on the masses of more than 2400 nuclei with $A=9-270$, we investigate the model dependence in the extraction of symmetry energy coefficient. We find that the extraction of the symmetry energy coefficients is strongly correlated with the forms of the Coulomb energy and the mass dependence of the symmetry energy coefficient adopted. The values of the extracted symmetry energy coefficients increase by about 2 MeV for heavy nuclei when the Coulomb correction term is involved. We obtain the bulk symmetry energy coefficient $S_0=28.26\pm1.3$ MeV and the surface-to-volume ratio $κ=1.26\pm 0.25 $ MeV if assuming the mass dependence of symmetry energy coefficient $a_{\rm sym}(A)=S_0(1-κ/A^{1/3})$, and $S_0=32.80\pm1.7$ MeV, $κ=2.82\pm0.57$ MeV when $a_{\rm sym}(A)=S_0 (1+κ/A^{1/3})^{-1}$ is adopted.

nucl-th↗