arXiv subjects

Matthew M. Hong

Publications and source records attributed to Matthew M. Hong.

4 recordsLinked to original sources

arXiv2026

Towards a Doubly Efficient IP=PSPACE

We show that every language in PSPACE decidable by a Turing machine in time $T(n)=n^{O(\log n)}$ admits a doubly efficient interactive proof system: the prover runs in time polynomial in T(n), and the verifier runs in time polynomial in n. This extends the best previously known regime for such proof systems from $T(n)=n^{O(\sqrt{\log n / \log\log n})}$, established by Berger, Goyal, Hong, and Kalai (FOCS 2025), to $T(n)=n^{O(\log n)}$. Beyond improving the range of T, our protocol is substantially simpler than previous doubly efficient proofs for time-bounded PSPACE. Earlier constructions proceed indirectly: they first build batch interactive proofs and then invoke them as a black box to obtain doubly efficient protocols. In contrast, we give a direct construction. This not only simplifies the proof but also points to a more promising route for future improvements.

cs.CC↗

arXiv2026

TMRL: Diffusion Timestep-Modulated Pretraining Enables Exploration for Efficient Policy Finetuning

Fine-tuning pre-trained robot policies with reinforcement learning (RL) often inherits the bottlenecks introduced by pre-training with behavioral cloning (BC), which produces narrow action distributions that lack the coverage necessary for downstream exploration. We present a unified framework that enables the exploration necessary to enable efficient robot policy finetuning by bridging BC pre-training and RL fine-tuning. Our pre-training method, Context-Smoothed Pre-training (CSP), injects forward-diffusion noise into policy inputs, creating a continuum between precise imitation and broad action coverage. We then fine-tune pre-trained policies via Timestep-Modulated Reinforcement Learning (TMRL), which trains the agent to dynamically adjust this conditioning during fine-tuning by modulating the diffusion timestep, granting explicit control over exploration. Integrating seamlessly with arbitrary policy inputs, e.g., states, 3D point clouds, or image-based VLA policies, we show that TMRL improves RL fine-tuning sample efficiency. Notably, TMRL enables successful real-world fine-tuning on complex manipulation tasks in under one hour. Videos and code available at https://weirdlabuw.github.io/tmrl/.

cs.RO↗

arXiv2025

Efficiently Batching Unambiguous Interactive Proofs

We show that if a language $L$ admits a public-coin unambiguous interactive proof (UIP) with round complexity $\ell$, where $a$ bits are communicated per round, then the batch language $L^{\otimes k}$, i.e. the set of $k$-tuples of statements all belonging to $L$, has an unambiguous interactive proof with round complexity $\ell\cdot\mathsf{polylog}(k)$, per-round communication of $a\cdot \ell\cdot\mathsf{polylog}(k) + \mathsf{poly}(\ell)$ bits, assuming the verifier in the $\mathsf{UIP}$ has depth bounded by $\mathsf{polylog}(k)$. Prior to this work, the best known batch $\mathsf{UIP}$ for $L^{\otimes{k}}$ required communication complexity at least $(\mathsf{poly}(a)\cdot k^{\epsilon} + k) \cdot \ell^{1/\epsilon}$ for any arbitrarily small constant $\epsilon>0$ (Reingold-Rothblum-Rothblum, STOC 2016). As a corollary of our result, we obtain a doubly efficient proof system, that is, a proof system whose proving overhead is polynomial in the time of the underlying computation, for any language computable in polynomial space and in time at most $n^{O\left(\sqrt{\frac{\log n}{\log\log n}}\right)}$. This expands the state of the art of doubly efficient proof systems: prior to our work, such systems were known for languages computable in polynomial space and in time $n^{({\log n})^\delta}$ for a small $\delta>0$ significantly smaller than $1/2$ (Reingold-Rothblum-Rothblum, STOC 2016).

cs.CC↗

arXiv2025

CLAM: Continuous Latent Action Models for Robot Learning from Unlabeled Demonstrations

Learning robot control policies from demonstrations typically requires action-labeled expert data, which is expensive to collect through teleoperation. We study a more practical setting in which expert demonstrations are available only as observation sequences without action labels, and only task-agnostic play data contains actions. We introduce continuous latent action models (CLAM), a framework that infers continuous latent actions between consecutive observations using self-supervised dynamics prediction. To ground these latent actions into executable motor commands, CLAM jointly trains an action decoder using a small amount of task-agnostic play data. We show that continuous latent actions combined with this joint training are essential for high-dimensional continuous control. Across DMControl locomotion, MetaWorld manipulation, and real-world WidowX robot tasks, CLAM improves average task success rates by 2-3x over prior latent-action baselines and approaches behavior cloning trained with privileged expert action labels. Our results demonstrate that effective robot policies can be learned from unlabeled demonstrations and deployed on real hardware without collecting expert action-labeled data. Videos and code are available at clamrobot.github.io.

cs.RO↗