Joint Empirical Coordination of Source and Channel

Treust, Maël Le

doi:10.1109/tit.2017.2714682

Cited by 19 publications

(16 citation statements)

References 74 publications

(148 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This chapter aims at describing a few special instances of a general problem which has been addressed in several recent works [2], [3], [4], [5], [6], [7], [8].…”

Section: General Problem Formulationmentioning

confidence: 99%

Information Constraints in Multiple Agent Problems with I.I.D. States

Lasaulce

Tarbouriech

2018

Control Subject to Computational and Communication Constraints

View full text Add to dashboard Cite

In this chapter we describe several recent results on the problem of coordination among agents when they have partial information about a state which affects their utility, payoff, or reward function. The state is not controlled and rather evolves according to an independent and identically distributed (i.i.d.) random process. This random process might represent various phenomena. In control, it may represent a perturbation or model uncertainty. In the context of smart grids, it may represent a forecasting noise [1]. In wireless communications, it may represent the state of the global communication channel. The approach used is to exploit Shannon theory to characterize the achievable long-term utility region. Two scenarios are described. In the first scenario, the number of agents is arbitrary and the agents have causal knowledge about the state. In the second scenario, there are only two agents and the agents have some knowledge about the future of the state, making its knowledge non-causal. Chapter overviewThis chapter concerns the problem of coordination among agents. Technically, the problem is as follows. We consider a set of K ≥ 2 agents. Agent k has a utility, payoff, or reward function u k (x 0 , x 1 , ..., x K ) where x k , k ≥ 1, is the action of Agent k while x 0 is the action of an agent called Nature. The Nature's actions correspond to the system state and is assumed to be non-controlled; more precisely, Nature corresponds to an independent and identically distributed (i.i.d.) random process. The problem studied in this chapter is to characterize the long-term utility region where σ k = (σ k,t ) t≥1 is a sequence of functions which represent the strategy of Agent k, x k (t) is the action chosen by Agent k at time or stage t ≥ 1, t being the time or stage index; concerning notations, as far as random variables are concerned, capital letters will stand for random variables whereas, small letters will stand for realizations. Note that, implicitly, we assume sufficient conditions (such as utility boundedness) under which the above limit exists. The functions σ k,t , k ∈ {1, ..., K}, map the available knowledge to the action of the considered agent. The available knowledge depends on the information assumptions made (e.g., the knowledge of the state can be causal or non-causal). We will distinguish between two scenarios. In the first scenario, agents are assumed to have some causal knowledge (in the wide sense) about the state whereas, in the second scenario non-causal knowledge (i.e., some knowledge about the future) about the state is assumed. The second scenario is definitely the most difficult one technically, which is why only two agents will be assumed. Remarkably, the long-term utility region, whenever available, can be characterized in terms of elegant information constraints. For instance, in the scenario of non-causal state information, determining the long-term utility region amounts to solving a convex optimization problem whose non-trivial constraints are the derived information-the...

show abstract

“…This chapter aims at describing a few special instances of a general problem which has been addressed in several recent works [2], [3], [4], [5], [6], [7], [8].…”

Section: General Problem Formulationmentioning

confidence: 99%

Information Constraints in Multiple Agent Problems with I.I.D. States

Lasaulce

Tarbouriech

2018

Control Subject to Computational and Communication Constraints

View full text Add to dashboard Cite

show abstract

“…• In [51, Theorem IV.2], the author shows that the sets Q 0 and Q 2 correspond to the target probability distributions Q(u, z, w) × Q(v|z, w) that are achievable for the problem of empirical coordination, see also [26], [28]. As noticed in [56] and [57], the tool of Empirical Coordination allows us to characterize the "core of the decoder's knowledge", that captures what the decoder can infer about all the random variables of the problem.…”

Section: Definition Iii1 (Target Distributions)mentioning

confidence: 99%

“…In their model, the sender observes an July 16, 2018 DRAFT infinite source of information and communicates with the receiver through a perfect channel of fixed alphabet. This result was later refined by Cuff in [25] and referred to as the "coordination problem" in several articles [26], [27], [28], [29]. In the literature of Information Theory, the Bayesian persuasion game is investigated in [30] for Gaussian source and channel with Crawford-Sobel's quadratic cost functions.…”

Section: Introductionmentioning

confidence: 99%

Information-Theoretic Limits of Strategic Communication

Treust

Tomala

2018

SSRN Journal

View full text Add to dashboard Cite

In this article, we investigate strategic information transmission over a noisy channel. This problem has been widely investigated in Economics, when the communication channel is perfect. Unlike in Information Theory, both encoder and decoder have distinct objectives and choose their encoding and decoding strategies accordingly. This approach radically differs from the conventional Communication paradigm, which assumes transmitters are of two types: either they have a common goal, or they act as opponent, e.g. jammer, eavesdropper. We formulate a point-to-point source-channel coding problem with state information, in which the encoder and the decoder choose their respective encoding and decoding strategies in order to maximize their long-run utility functions. This strategic coding problem is at the interplay between Wyner-Ziv's scenario and the Bayesian persuasion game of Kamenica-Gentzkow. We characterize a single-letter solution and we relate it to the previous results by using the concavification method. This confirms the benefit of sending encoded data bits even if the decoding process is not supervised, e.g. when the decoder is an autonomous device. Our solution has two interesting features: it might be optimal not to use all channel resources; the informational content impacts the encoding process, since utility functions capture preferences on source symbols. * Maël Le Treust gratefully acknowledges the support of the Labex MME-DII (ANR11-LBX-0023-01) and the DIM-RFSI of the Région Île-de-France, under grant EX032965.† Tristan Tomala gratefully acknowledges the support of the HEC foundation and ANR/Investissements d'Avenir under grant ANR-11-IDEX-0003/Labex Ecodec/ANR-11-LABX-0047.

show abstract

“…Strong coordination in networks was first studied over error free links [3] and later extended to noisy communication links [11]. In the latter setting, the signals that are transmitted and received over the physical channel become a part of what can be observed, and one can therefore coordinate the actions of the devices with their communication signals [13,14]. From a security standpoint, this joint coordination of actions and signals allows one to control the information about the devices' actions that may be inferred from the observations of the communication signals.…”

Section: Introductionmentioning

confidence: 99%

Strong Coordination of Signals and Actions Over Noisy Channels With Two-Sided State Information

Cervia

Luzzi

Treust

et al. 2020

IEEE Trans. Inform. Theory

Self Cite

View full text Add to dashboard Cite

We consider a network of two nodes separated by a noisy channel with two-sided state information, in which the input and output signals have to be coordinated with the source and its reconstruction. In the case of non-causal encoding and decoding, we propose a joint source-channel coding scheme and develop inner and outer bounds for the strong coordination region. While the inner and outer bounds do not match in general, we provide a complete characterization of the strong coordination region in three particular cases: i) when the channel is perfect; ii) when the decoder is lossless; and iii) when the random variables of the channel are independent from the random variables of the source. Through the study of these special cases, we prove that the separation principle does not hold for joint source-channel strong coordination. Finally, in the absence of state information, we show that polar codes achieve the best known inner bound for the strong coordination region, which therefore offers a constructive alternative to random binning and coding proofs.

show abstract

Joint Empirical Coordination of Source and Channel

Cited by 19 publications

References 74 publications

Information Constraints in Multiple Agent Problems with I.I.D. States

Information Constraints in Multiple Agent Problems with I.I.D. States

Information-Theoretic Limits of Strategic Communication

Strong Coordination of Signals and Actions Over Noisy Channels With Two-Sided State Information

Contact Info

Product

Resources

About