arXiv subjects

Selma Belhadj Amor

Publications and source records attributed to Selma Belhadj Amor.

7 recordsLinked to original sources

arXiv2024

From English-Centric to Effective Bilingual: LLMs with Custom Tokenizers for Underrepresented Languages

In this paper, we propose a model-agnostic cost-effective approach to developing bilingual base large language models (LLMs) to support English and any target language. The method includes vocabulary expansion, initialization of new embeddings, model training and evaluation. We performed our experiments with three languages, each using a non-Latin script - Ukrainian, Arabic, and Georgian. Our approach demonstrates improved language performance while reducing computational costs. It mitigates the disproportionate penalization of underrepresented languages, promoting fairness and minimizing adverse phenomena such as code-switching and broken grammar. Additionally, we introduce new metrics to evaluate language quality, revealing that vocabulary size significantly impacts the quality of generated text.

cs.CL↗

arXiv2018

Distributed Hypothesis Testing with Privacy Constraints

We revisit the distributed hypothesis testing (or hypothesis testing with communication constraints) problem from the viewpoint of privacy. Instead of observing the raw data directly, the transmitter observes a sanitized or randomized version of it. We impose an upper bound on the mutual information between the raw and randomized data. Under this scenario, the receiver, which is also provided with side information, is required to make a decision on whether the null or alternative hypothesis is in effect. We first provide a general lower bound on the type-II exponent for an arbitrary pair of hypotheses. Next, we show that if the distribution under the alternative hypothesis is the product of the marginals of the distribution under the null (i.e., testing against independence), then the exponent is known exactly. Moreover, we show that the strong converse property holds. Using ideas from Euclidean information theory, we also provide an approximate expression for the exponent when the communication rate is low and the privacy level is high. Finally, we illustrate our results with a binary and a Gaussian example.

cs.IT↗

arXiv2017

Decentralized Simultaneous Information and Energy Transmission in K-User Multiple Access Channels

In this paper, the fundamental limits of decentralized simultaneous information and energy transmission in the $K$-user Gaussian multiple access channel (G-MAC), with an arbitrary $K \geqslant 2$ and one non-colocated energy harvester (EH), are fully characterized. The objective of the transmitters is twofold. First, they aim to reliably communicate their message indices to the receiver; and second, to harvest energy at the EH at a rate not less than a minimum rate requirement $b$. The information rates $R_1,\dots,R_K$, in bits per channel use, are measured at the receiver and the energy rate $B$ is measured at an EH. Stability is considered in the sense of an $η$-Nash equilibrium ($η$-NE), with $η> 0$. The main result is a full characterization of the $η$-NE information-energy region, i.e., the set of information-energy rate tuples $(R_1,\dots,R_K,B)$ that are achievable and stable in the G-MAC when: $(a)$ all the transmitters autonomously and independently tune their own transmit configurations seeking to maximize their own information transmission rates $R_1,\dots, R_K$; and $(b)$ all the transmitters jointly guarantee an energy transmission rate $B$ at the EH, such that $B \geqslant b$. Therefore, any rate tuple outside the $η$-NE region is not stable as there always exists at least one transmitter able to increase by at least $η$ bits per channel use its own information transmission rate by updating its own transmit configuration.

cs.IT↗

arXiv2017

Feedback Enhances Simultaneous Wireless Information and Energy Transmission in Multiple Access Channels

In this report, the fundamental limits of simultaneous information and energy transmission in the two-user Gaussian multiple access channel (G-MAC) with and without feedback are fully characterized. More specifically, all the achievable information and energy transmission rates (in bits per channel use and energy-units per channel use, respectively) are identified. Furthermore, the fundamental limits on the individual and sum- rates given a minimum energy rate ensured at an energy harvester are also characterized. In the case without feedback, an achievability scheme based on power-splitting and successive interference cancellation is shown to be optimal. Alternatively, in the case with feedback (G-MAC-F), a simple yet optimal achievability scheme based on power-splitting and Ozarow's capacity achieving scheme is presented. Finally, the energy transmission enhancement induced by the use of feedback is quantified. Feedback can at most double the energy transmission rate at high SNRs when the information transmission sum-rate is kept fixed at the sum-capacity of the G-MAC, but it has no effect at very low SNRs.

cs.IT↗

arXiv2016

Fundamental Limits of Simultaneous Energy and Information Transmission

In this paper, existing results regarding the fundamental limits of simultaneous energy and information transmission in wireless networks are reviewed. In point-to-point channels, the fundamental limits on the information rate given a minimum energy rate constraint are fully characterized by the notion of information-energy capacity function introduced by Varshney in 2008. In a centralized multi-user channel, the fundamental limits on the information rates given a minimum energy rate constraint are described by the notion of information-energy capacity region introduced by Fouladgar and Simeone in 2012. Alternatively, in a decentralized multi-user channel, the authors have recently introduced the notion of information-energy Nash region to describe these fundamental limits. All these fundamental limits reveal the intrinsic trade-off between the two conflicting tasks of information and energy transmission.

cs.IT↗

arXiv2015

A New Coding Scheme for Discrete Memoryless MACs with Common Rate-Limited Feedback

We propose a new coding scheme for the discrete memoryless two-user multi-access channel (MAC) with rate-limited feedback. Our scheme combines ideas from the Venkataramanan-Pradhan scheme for perfect feedback with ideas from the Shaviv-Steinberg scheme for rate-limited feedback. Our achievable region includes the Shaviv-Steinberg achievable region and this inclusion can be strict. For general MACs and for sufficiently large feedback rates, our scheme outperforms the Shaviv-Steinberg scheme as it achieves the same rate region as the Venkataramanan-Pradhan scheme for perfect feedback (which cannot be achieved by the Shaviv-Steinberg scheme). Furthermore, we numerically evaluate our achievable region with a specific (Gaussian) choice of random variables for the memoryless two-user Gaussian MAC. Our simulation results show that for some parameters of the Gaussian MAC and the feedback rate, our scheme achieves a strictly larger sum-rate than the Shaviv-Steinberg scheme.

cs.IT↗

arXiv2014

MIMO MAC-BC Duality with Linear-Feedback Coding Schemes

We show that for the multi-antenna Gaussian multi-access channel (MAC) and broadcast channel (BC) with perfect feedback, the rate regions achieved by linear-feedback coding schemes (called linear-feedback capacity regions) coincide when the same total input-power constraint is imposed on both channels and when the MAC channel matrices are the transposes of the BC channel matrices. Such a pair of MAC and BC is called dual. We also identify sub-classes of linear-feedback coding schemes that achieve the linear-feedback capacity regions of these two channels and present multi-letter expressions for the linear-feedback capacity regions. Moreover, within the two sub-classes of coding schemes that achieve the linear-feedback capacity regions for a given MAC and its dual BC, we identify for each MAC scheme a BC scheme and for each BC scheme a MAC scheme so that the two schemes have same total input power and achieve the same rate regions.

cs.IT↗