Chenheng Zhang

The cost of tracking changing comparators in exp-concave online learning is reduced by a factor of $d^{1/3}$, resolving the dimension's optimal dependence.

Up to 47% of reported fine-tuning gains in tool retrieval are an evaluation illusion caused by benchmarks treating functionally equivalent tools as retrieval failures.

UniMem achieves a remarkable 4.0 EM point improvement in task execution fidelity by seamlessly integrating episodic and parametric memory without the need for explicit task labels.