Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2511.19947

Towards Edge General Intelligence: Knowledge Distillation for Mobile Agentic AI

Yuxuan Wu·Linghan Ma·Ruichen Zhang·Yinqiu Liu·Dusit Niyato·Shunpu Tang·Zehui Xiong·Zhu Han·Zhaohui Yang·Kaibin Huang·Zhaoyang Zhang·Kai-Kit Wong
View original source ↗Read source document ↗Export metadata ↓

Abstract

Edge General Intelligence (EGI) represents a paradigm shift in mobile edge computing, where intelligent agents operate autonomously in dynamic, resource-constrained environments. However, the deployment of advanced agentic AI models on mobile and edge devices faces significant challenges due to limited computation, energy, and storage resources. To address these constraints, this survey investigates the integration of Knowledge Distillation (KD) into EGI, positioning KD as a key enabler for efficient, communication-aware, and scalable intelligence at the wireless edge. In particular, we emphasize KD techniques specifically designed for wireless communication and mobile networking, such as channel-aware self-distillation, cross-model Channel State Information (CSI) feedback distillation, and robust modulation/classification distillation. Furthermore, we review novel architectures natively suited for KD and edge deployment, such as Mamba, RWKV (Receptance, Weight, Key, Value) and Cross-Architecture distillation, which enhance generalization capabilities. Subsequently, we examine diverse applications in which KD-driven architectures enable EGI across vision, speech, and multimodal tasks. Finally, we highlight the key challenges and future directions for KD in EGI. This survey aims to provide a comprehensive reference for researchers exploring KD-driven frameworks for mobile agentic AI in the era of EGI.

Explore related subjects

cs.ITeess.SPmath.IT

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Yuxuan Wu, Linghan Ma, Ruichen Zhang, Yinqiu Liu, Dusit Niyato, Shunpu Tang, Zehui Xiong, Zhu Han, Zhaohui Yang, Kaibin Huang, Zhaoyang Zhang, Kai-Kit Wong. 2025-11-25. Towards Edge General Intelligence: Knowledge Distillation for Mobile Agentic AI. https://arxiv.org/abs/2511.19947

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Efficient and rate-optimal list-decoding in the presence of minimal feedback

Given a channel with length-$n$ inputs and outputs over the alphabet $\{0,1,\ldots,q-1\}$, and of which a fraction $\varrho \in (0,1-1/q)$ of symbols can be arbitrarily corrupted by an adversary, a fundamental problem is that of communicating at rates close to the information-theoretically optimal values, while ensuring the receiver can infer that the transmitter's message is from a ``small" set. While the existence of such codes is known, and constructions with computationally tractable encoding/decoding procedures are known for large $q$, we provide the first schemes that attain this performance for any $q \geq 2$, as long as low-rate feedback (asymptotically negligible relative to the number of transmissions) from the receiver to the transmitter is available. For any sufficiently small $\varepsilon > 0$ and $\varrho \in (1-{1}/{q}-Θ(\sqrt{\varepsilon}))$ our minimal feedback scheme has the following parameters: Rate $1-H_q(\varrho) - \varepsilon$ (i.e., $\varepsilon$-close to information-theoretically optimal -- here $H_q(\varrho)$ is the $q$-ary entropy function), list-size $\exp\left(\mathcal{O}\left(\varepsilon^{-3/2}\log^2(1/\varepsilon)\right)\right)$, computational complexity of encoding/decoding $n^{\mathcal{O}(\varepsilon^{-1}\log(1/\varepsilon))}$, storage complexity $\mathcal{O}(n^{η+1}\log n)$ for a code design parameter $η>1$ that trades off storage complexity with the probability of error. The error probability is $\mathcal{O}(n^{-η})$, and the (vanishing) feedback rate is $\mathcal{O}({1}/{\sqrt{\log(n)}})$. Our full-feedback scheme has zero probability of error and minimal storage complexity, while the other parameters are the same as the vanishing rate feedback scheme.

cs.IT↗
arXiv2026

On Cost-Aware Designs for Sequential Hypothesis Testing

We introduce Cost-Aware (CA) Sequential Hypothesis Testing (CASHT), in which an active decision-maker selects sensing actions with different, random costs to identify the true hypothesis under an average-error constraint $δ$, while minimizing the expected total cost (rather than the number of samples). For fixed costs, we prove that the optimal expected total cost scales as $Θ(\log(1/δ))$, and is achievable by Multihypothesis Sequential Probability Ratio Test-based procedures. We show that the CA design principle is to maximize the ratio of expected information gain to expected cost under the policy-induced action distribution. Guided by this principle, we adapt two classic policies to the CA setting and establish their asymptotic optimality. We then treat random costs under two revelation models: ex-post, where costs are disclosed only after a sample is obtained, and the cost-error tradeoff coincides with the fixed-cost case, and ex-ante, where costs accrue before acquisition, and the decision maker may cancel an action mid-operation. For the ex-ante model, we characterize when cancellation lowers the total cost and analyze several cost distributions in detail. Simulations confirm our findings that the CA variants consistently reduce total cost relative to their classic counterparts, and when action cancellation helps or hurts.

cs.IT↗
arXiv2026

All you need is log

How different are several probability distributions from one another? For two distributions the standard answer is the family of Rényi divergences, singled out by two natural requirements: processing the data never makes distributions easier to tell apart, and independent repetitions add. Many problems in learning and statistics compare more than two distributions at once, such as testing among several hypotheses or bounding generalization against several priors. The same two requirements leave one kind of building block, built on a coincidence probability: how unlikely it is that independent samples, one from each distribution, all show the same empirical distribution. The logarithm is forced because repetitions add, which is already visible for a single experiment repeated. This characterization is known in greater generality, and this paper is about the meaning of its building blocks. On a finite alphabet, each building block indexed by a rational point of the simplex is the exponential rate of that coincidence as the samples grow in fixed proportions. Each is also the limiting free energy of Bayesian inference over distributions. At any amount of data, the free energy of the posterior is the coincidence measure plus two costs: the expected distance from a posterior draw to the most likely distribution, and the information gained per unit of data. Both costs vanish as data accumulate. When the comparison is conditioned on side information, every kind of building block has a conditional counterpart, and the coincidence ones alone do not suffice.

cs.IT↗