arXiv subjects

Tomohiro Inoue

Publications and source records attributed to Tomohiro Inoue.

3 recordsLinked to original sources

arXiv2025

Distinct Profiles of Run-to-Run Score Reliability and Expert-Panel Alignment Across Four LLM Evaluators of Simulated Japanese-Language AI-to-AI Counseling

Large language models (LLMs) increasingly evaluate generated dialogue, but repeatable scores do not necessarily align with professional judgment. This observational fixed-benchmark study compared four configured LLM evaluator systems (GPT-5.5, Gemini 3.5 Flash, Claude Opus 4.8, and Fable 5) with aggregated ratings from 15 counseling experts on 18 complete simulated AI-to-AI counseling sessions conducted in Japanese. The sessions represented three counselor conditions across six prespecified client profiles. Each system scored every transcript three times on four motivational interviewing-informed dimensions and overall quality. All four systems assigned higher scores than the expert panel for softening sustain talk and overall quality, although differences varied across systems and constructs. Single-run intraclass correlation coefficients ranged from .33 to .96, showing that high run-to-run reliability did not ensure closer expert-panel alignment. Claude Opus 4.8 had the smallest mean absolute difference, whereas Fable 5 had an intermediate difference. In a secondary benchmark analysis, GPT-4-turbo sessions generated with the Structured Multi-step Dialogue Prompt received higher expert ratings than sessions generated by the same model with a minimal instruction for cultivating change talk, partnership, empathy, and overall quality; the softening sustain talk contrast remained uncertain. The fixed benchmark contained one session per counselor-condition-by-profile cell, so inference concerns these sessions rather than all possible stochastic regenerations. Run-to-run reliability, expert-panel alignment, and condition discrimination are separate properties of automated counseling evaluation. This benchmark supports construct-level assessment of LLM evaluators against professional judgment.

cs.CL↗

arXiv2021

Pruning Randomly Initialized Neural Networks with Iterative Randomization

Pruning the weights of randomly initialized neural networks plays an important role in the context of lottery ticket hypothesis. Ramanujan et al. (2020) empirically showed that only pruning the weights can achieve remarkable performance instead of optimizing the weight values. However, to achieve the same level of performance as the weight optimization, the pruning approach requires more parameters in the networks before pruning and thus more memory space. To overcome this parameter inefficiency, we introduce a novel framework to prune randomly initialized neural networks with iteratively randomizing weight values (IteRand). Theoretically, we prove an approximation theorem in our framework, which indicates that the randomizing operations are provably effective to reduce the required number of the parameters. We also empirically demonstrate the parameter efficiency in multiple experiments on CIFAR-10 and ImageNet. The code is available at: https://github.com/dchiji-ntt/iterand

cs.LG↗

arXiv2019

Topological band structure of surface acoustic waves on a periodically corrugated surface

Surface acoustic waves (SAWs) are elastic waves localized on a surface of an elastic body. We theoretically study topological edge modes of SAWs for a corrugated surface. We introduce a corrugation forming a triangular lattice on the surface of an elastic body. We treat the corrugation as a perturbation, and construct eigenmodes on a corrugated surface by superposing those for the flat surface at wavevectors which are mutually different by reciprocal lattice vectors. We thereby show emergence of Dirac cones at the $K$ and $K'$ points analytically. Moreover, by breaking the time-reversal symmetry, we show that the Dirac cones open a gap, and that the Chern number for the lowest band has a nonzero value. It means existence of topological chiral edge modes of SAWs in the gap.

cond-mat.mes-hall↗