arXiv subjects

Zied Ben Houidi

Publications and source records attributed to Zied Ben Houidi.

16 recordsLinked to original sources

arXiv2026

Disentangling Self-Distillation: Measuring and Modeling Acquisition and Retention

Self-distillation with privileged context adapts a language model from demonstrations by letting the model, once conditioned on a reference response, teach its context-free copy token by token. Our taxonomy reveals existing methods differ along three entangled axes: (i) the rollout source (student or teacher), (ii) the teacher coupling (frozen, or an exponential moving average of the student at some coupling rate) and (iii) the KL direction (reverse or forward), yet these axes are usually studied in fixed combinations and have led to conflicting conclusions. We formalize a unifying framework to encompass all self-distillation methods vs classic supervised fine-tuning: we train every combination of the three axes, on Qwen2.5-7B and Ministral-3-3B across ordinary and contradictory tasks, totaling 1,200 adaptation runs, to systematically investigate the impact of the above axes. We propose a controlled model of the same objective to explain the resulting acquisition-retention trade-offs. We find that (i) the rollout source matters mostly where the task contradicts the pretrained behavior: there teacher rollouts raise acquisition well above what student rollouts achieve, with almost no change in retention; (ii) the teacher coupling changes acquisition most, on every task: acquisition rises with the coupling rate, then falls past a task-specific rate; (iii) switching the KL direction costs retention in one model but not the other so which axis to tune first depends on the model. The controlled model reproduces the three trends.

cs.AI↗

arXiv2025

In Praise of Stubbornness: An Empirical Case for Cognitive-Dissonance Aware Continual Update of Knowledge in LLMs

Through systematic empirical investigation, we uncover a fundamental and concerning property of Large Language Models: while they can safely learn facts that don't contradict their knowledge, attempting to update facts with contradictory information triggers catastrophic corruption of unrelated knowledge. Unlike humans, who naturally resist contradictory information, these models indiscriminately accept contradictions, leading to devastating interference, destroying up to 80% of unrelated knowledge even when learning as few as 10-100 contradicting facts. To understand whether this interference could be mitigated through selective plasticity, we experiment with targeted network updates, distinguishing between previously used (stubborn) and rarely used (plastic) neurons. We uncover another asymmetry: while sparing frequently-used neurons significantly improves retention of existing knowledge for non-contradictory updates (98% vs 93% with standard updates), contradictory updates trigger catastrophic interference regardless of targeting strategy. This effect which persists across tested model scales (GPT-2 to GPT-J-6B), suggests a fundamental limitation in how neural networks handle contradictions. Finally, we demonstrate that contradictory information can be reliably detected (95%+ accuracy) using simple model features, offering a potential protective mechanism. These findings motivate new architectures that can, like humans, naturally resist contradictions rather than allowing destructive overwrites.

cs.CL↗

arXiv2025

Episodic Memories Generation and Evaluation Benchmark for Large Language Models

Episodic memory -- the ability to recall specific events grounded in time and space -- is a cornerstone of human cognition, enabling not only coherent storytelling, but also planning and decision-making. Despite their remarkable capabilities, Large Language Models (LLMs) lack a robust mechanism for episodic memory: we argue that integrating episodic memory capabilities into LLM is essential for advancing AI towards human-like cognition, increasing their potential to reason consistently and ground their output in real-world episodic events, hence avoiding confabulations. To address this challenge, we introduce a comprehensive framework to model and evaluate LLM episodic memory capabilities. Drawing inspiration from cognitive science, we develop a structured approach to represent episodic events, encapsulating temporal and spatial contexts, involved entities, and detailed descriptions. We synthesize a unique episodic memory benchmark, free from contamination, and release open source code and datasets to assess LLM performance across various recall and episodic reasoning tasks. Our evaluation of state-of-the-art models, including GPT-4 and Claude variants, Llama 3.1, and o1-mini, reveals that even the most advanced LLMs struggle with episodic memory tasks, particularly when dealing with multiple related events or complex spatio-temporal relationships -- even in contexts as short as 10k-100k tokens.

cs.CL↗

arXiv2024

Fine-grained Attention in Hierarchical Transformers for Tabular Time-series

Tabular data is ubiquitous in many real-life systems. In particular, time-dependent tabular data, where rows are chronologically related, is typically used for recording historical events, e.g., financial transactions, healthcare records, or stock history. Recently, hierarchical variants of the attention mechanism of transformer architectures have been used to model tabular time-series data. At first, rows (or columns) are encoded separately by computing attention between their fields. Subsequently, encoded rows (or columns) are attended to one another to model the entire tabular time-series. While efficient, this approach constrains the attention granularity and limits its ability to learn patterns at the field-level across separate rows, or columns. We take a first step to address this gap by proposing Fieldy, a fine-grained hierarchical model that contextualizes fields at both the row and column levels. We compare our proposal against state of the art models on regression and classification tasks using public tabular time-series datasets. Our results show that combining row-wise and column-wise attention improves performance without increasing model size. Code and data are available at https://github.com/raphaaal/fieldy.

cs.LG↗

arXiv2024

Generic Multi-modal Representation Learning for Network Traffic Analysis

Network traffic analysis is fundamental for network management, troubleshooting, and security. Tasks such as traffic classification, anomaly detection, and novelty discovery are fundamental for extracting operational information from network data and measurements. We witness the shift from deep packet inspection and basic machine learning to Deep Learning (DL) approaches where researchers define and test a custom DL architecture designed for each specific problem. We here advocate the need for a general DL architecture flexible enough to solve different traffic analysis tasks. We test this idea by proposing a DL architecture based on generic data adaptation modules, followed by an integration module that summarises the extracted information into a compact and rich intermediate representation (i.e. embeddings). The result is a flexible Multi-modal Autoencoder (MAE) pipeline that can solve different use cases. We demonstrate the architecture with traffic classification (TC) tasks since they allow us to quantitatively compare results with state-of-the-art solutions. However, we argue that the MAE architecture is generic and can be used to learn representations useful in multiple scenarios. On TC, the MAE performs on par or better than alternatives while avoiding cumbersome feature engineering, thus streamlining the adoption of DL solutions for traffic analysis.

cs.LG↗

arXiv2024

LogPrécis: Unleashing Language Models for Automated Malicious Log Analysis

The collection of security-related logs holds the key to understanding attack behaviors and diagnosing vulnerabilities. Still, their analysis remains a daunting challenge. Recently, Language Models (LMs) have demonstrated unmatched potential in understanding natural and programming languages. The question arises whether and how LMs could be also useful for security experts since their logs contain intrinsically confused and obfuscated information. In this paper, we systematically study how to benefit from the state-of-the-art in LM to automatically analyze text-like Unix shell attack logs. We present a thorough design methodology that leads to LogPrécis. It receives as input raw shell sessions and automatically identifies and assigns the attacker tactic to each portion of the session, i.e., unveiling the sequence of the attacker's goals. We demonstrate LogPrécis capability to support the analysis of two large datasets containing about 400,000 unique Unix shell attacks. LogPrécis reduces them into about 3,000 fingerprints, each grouping sessions with the same sequence of tactics. The abstraction it provides lets the analyst better understand attacks, identify fingerprints, detect novelty, link similar attacks, and track families and mutations. Overall, LogPrécis, released as open source, paves the way for better and more responsive defense against cyberattacks.

cs.CR↗

arXiv2023

User-aware WLAN Transmit Power Control in the Wild

In Wireless Local Area Networks (WLANs), Access point (AP) transmit power influences (i) received signal quality for users and thus user throughput, (ii) user association and thus load across APs and (iii) AP coverage ranges and thus interference in the network. Despite decades of academic research, transmit power levels are still, in practice, statically assigned to satisfy uniform coverage objectives. Yet each network comes with its unique distribution of users in space, calling for a power control that adapts to users' probabilities of presence, for example, placing the areas with higher interference probabilities where user density is the lowest. Although nice on paper, putting this simple idea in practice comes with a number of challenges, with gains that are difficult to estimate, if any at all. This paper is the first to address these challenges and evaluate in a production network serving thousands of daily users the benefits of a user-aware transmit power control system. Along the way, we contribute a novel approach to reason about user densities of presence from historical IEEE 802.11k data, as well as a new machine learning approach to impute missing signal-strength measurements. Results of a thorough experimental campaign show feasibility and quantify the gains: compared to state-of-the-art solutions, the new system can increase the median signal strength by 15dBm, while decreasing airtime interference at the same time. This comes at an affordable cost of a 5dBm decrease in uplink signal due to lack of terminal cooperation.

cs.NI↗

arXiv2022

Cross-network transferable neural models for WLAN interference estimation

Airtime interference is a key performance indicator for WLANs, measuring, for a given time period, the percentage of time during which a node is forced to wait for other transmissions before to transmitting or receiving. Being able to accurately estimate interference resulting from a given state change (e.g., channel, bandwidth, power) would allow a better control of WLAN resources, assessing the impact of a given configuration before actually implementing it. In this paper, we adopt a principled approach to interference estimation in WLANs. We first use real data to characterize the factors that impact it, and derive a set of relevant synthetic workloads for a controlled comparison of various deep learning architectures in terms of accuracy, generalization and robustness to outlier data. We find, unsurprisingly, that Graph Convolutional Networks (GCNs) yield the best performance overall, leveraging the graph structure inherent to campus WLANs. We notice that, unlike e.g. LSTMs, they struggle to learn the behavior of specific nodes, unless given the node indexes in addition. We finally verify GCN model generalization capabilities, by applying trained models on operational deployments unseen at training time.

cs.NI↗

arXiv2022

Neural language models for network configuration: Opportunities and reality check

Boosted by deep learning, natural language processing (NLP) techniques have recently seen spectacular progress, mainly fueled by breakthroughs both in representation learning with word embeddings (e.g. word2vec) as well as novel architectures (e.g. transformers). This success quickly invited researchers to explore the use of NLP techniques to other fields, such as computer programming languages, with the promise to automate tasks in software programming (bug detection, code synthesis, code repair, cross language translation etc.). By extension, NLP has potential for application to network configuration languages as well, for instance considering tasks such as network configuration verification, synthesis, and cross-vendor translation. In this paper, we survey recent advances in deep learning applied to programming languages, for the purpose of code verification, synthesis and translation: in particularly, we review their training requirements and expected performance, and qualitatively assess whether similar techniques can benefit corresponding use-cases in networking.

cs.NI↗

arXiv2022

From Design to Deployment of Zero-touch Deep Reinforcement Learning WLANs

Machine learning (ML) is increasingly used to automate networking tasks, in a paradigm known as zero-touch network and service management (ZSM). In particular, Deep Reinforcement Learning (DRL) techniques have recently gathered much attention for their ability to learn taking complex decisions in different fields. In the ZSM context, DRL is an appealing candidate for tasks such as dynamic resource allocation, that is generally formulated as hard optimization problems. At the same time, successful training and deployment of DRL agents in real-world scenarios faces a number of challenges that we outline and address in this paper. Tackling the case of Wireless Local Area Network (WLAN) radio resource management, we report guidelines that extend to other usecases and more general contexts.

cs.NI↗

arXiv2022

Neural combinatorial optimization beyond the TSP: Existing architectures under-represent graph structure

Recent years have witnessed the promise that reinforcement learning, coupled with Graph Neural Network (GNN) architectures, could learn to solve hard combinatorial optimization problems: given raw input data and an evaluator to guide the process, the idea is to automatically learn a policy able to return feasible and high-quality outputs. Recent work have shown promising results but the latter were mainly evaluated on the travelling salesman problem (TSP) and similar abstract variants such as Split Delivery Vehicle Routing Problem (SDVRP). In this paper, we analyze how and whether recent neural architectures can be applied to graph problems of practical importance. We thus set out to systematically "transfer" these architectures to the Power and Channel Allocation Problem (PCAP), which has practical relevance for, e.g., radio resource allocation in wireless networks. Our experimental results suggest that existing architectures (i) are still incapable of capturing graph structural features and (ii) are not suitable for problems where the actions on the graph change the graph attributes. On a positive note, we show that augmenting the structural representation of problems with Distance Encoding is a promising step towards the still-ambitious goal of learning multi-purpose autonomous solvers.

cs.AI↗

arXiv2021

The backpropagation-based recollection hypothesis: Backpropagated action potentials mediate recall, imagination, language understanding and naming

Ever since the advent of the neuron doctrine more than a century ago, information processing in the brain is widely believed to mainly follow the forward pre to post-synaptic neurons direction. Challenging this prevalent view, in this paper, we emit the backpropagation-based recollection hypothesis as follows: weak and fast fading Action Potentials following the (highest weight) post to pre-synaptic backward pathways, mediate explicit cue-based memory recall. This includes also the tasks of imagination, future episodic thinking, language understanding and associating names to various stimuli. These signals originate in highly invariant neurons, which uniquely respond to some specific stimuli (e.g. image of a cat). They then travel backwards to reactivate the same populations of neurons that uniquely respond to this specific stimuli during perception, thus recreating "offline" an experience that is similar. After stating our hypothesis in details, we challenge its assumptions through a thorough literature review. We find abundant evidence that supports most of its assumptions, including the existence of backpropagating signals that have interesting properties. We then leverage simulations based on existing spiking neural network models with STDP learning to show the computational feasibility of using such a mechanism to map the image of an object to its name with the same high accuracy as a state of the art machine learning classifier. Although not yet a theory, we believe this hypothesis presents a paradigm shift that is worth further investigating: it opens the way, among others, to new interpretations of language acquisition and understanding, the interplay between memories encoding and retrieval, as well as reconciling the apparently opposed views between sparse coding and distributed representations.

q-bio.NC↗

arXiv2021

You, the Web and Your Device: Longitudinal Characterization of Browsing Habits

Understanding how people interact with the web is key for a variety of applications, e.g., from the design of effective web pages to the definition of successful online marketing campaigns. Browsing behavior has been traditionally represented and studied by means of clickstreams, i.e., graphs whose vertices are web pages, and edges are the paths followed by users. Obtaining large and representative data to extract clickstreams is however challenging. The evolution of the web questions whether browsing behavior is changing and, by consequence, whether properties of clickstreams are changing. This paper presents a longitudinal study of clickstreams in from 2013 to 2016. We evaluate an anonymized dataset of HTTP traces captured in a large ISP, where thousands of households are connected. We first propose a methodology to identify actual URLs requested by users from the massive set of requests automatically fired by browsers when rendering web pages. Then, we characterize web usage patterns and clickstreams, taking into account both the temporal evolution and the impact of the device used to explore the web. Our analyses precisely quantify various aspects of clickstreams and uncover interesting patterns, such as the typical short paths followed by people while navigating the web, the fast increasing trend in browsing from mobile devices and the different roles of search engines and social networks in promoting content. Finally, we contribute a dataset of anonymized clickstreams to the community to foster new studies (anonymized clickstreams are available to the public at http://bigdata.polito.it/clickstream).

cs.CY↗

arXiv2020

White Paper on Crowdsourced Network and QoE Measurements -- Definitions, Use Cases and Challenges

This white paper is the outcome of the Würzburg seminar on "Crowdsourced Network and QoE Measurements" which took place from 25-26 September 2019 in Würzburg, Germany. International experts were invited from industry and academia. They are well known in their communities, having different backgrounds in crowdsourcing, mobile networks, network measurements, network performance, Quality of Service (QoS), and Quality of Experience (QoE). The discussions in the seminar focused on how crowdsourcing will support vendors, operators, and regulators to determine the Quality of Experience in new 5G networks that enable various new applications and network architectures. As a result of the discussions, the need for a white paper manifested, with the goal of providing a scientific discussion of the terms "crowdsourced network measurements" and "crowdsourced QoE measurements", describing relevant use cases for such crowdsourced data, and its underlying challenges. During the seminar, those main topics were identified, intensively discussed in break-out groups, and brought back into the plenum several times. The outcome of the seminar is this white paper at hand which is - to our knowledge - the first one covering the topic of crowdsourced network and QoE measurements.

cs.NI↗

arXiv2016

WeBrowse: Mining HTTP logs online for network-based content recommendation

A powerful means to help users discover new content in the overwhelming amount of information available today is sharing in online communities such as social networks or crowdsourced platforms. This means comes short in the case of what we call communities of a place: people who study, live or work at the same place. Such people often share common interests but either do not know each other or fail to actively engage in submitting and relaying information. To counter this effect, we propose passive crowdsourced content discovery, an approach that leverages the passive observation of web-clicks as an indication of users' interest in a piece of content. We design, implement, and evaluate WeBrowse , a passive crowdsourced system which requires no active user engagement to promote interesting content to users of a community of a place. Instead, it extracts the URLs users visit from traffic traversing a network link to identify popular and interesting pieces of information. We first prototype WeBrowse and evaluate it using both ground-truths and real traces from a large European Internet Service Provider. Then, we deploy WeBrowse in a campus of 15,000 users, and in a neighborhood. Evaluation based on our deployments shows the feasibility of our approach. The majority of WeBrowse's users welcome the quality of content it promotes. Finally, our analysis of popular topics across different communities confirms that users in the same community of a place share common interests, compared to users from different communities, thus confirming the promise of WeBrowse's approach.

cs.HC↗

arXiv2015

From Rough Consensus to Automated Reasoning

Unwritten languages today often have no official grammar, and are rather governed by "unspoken rules". Similarly, we argue that the young discipline of networking is still a practice that lacks a deep understanding of the rules that govern it. This situation results in a tremendous loss of time and efforts. First, since the rules are unspoken, they are not systematically reused. Second, since there is no grammar, it is impossible to assert if a sentence is correct. Comparing two networking approaches or solutions is sometimes a synonym of endless religious debates. Drawing the proper conclusion from this claim, we advocate that networking research should spend more efforts on better understanding its rules as a first step to automatically reuse them. To illustrate our claim, we focus in this paper on one specific networking problem, and show how different instances of this same problem were solved in parallel, resulting in different solutions, and no explicit knowledge reuse.

cs.NI↗