arXiv:2511.05297v2 Announce Type: replace Abstract: Digital Adoption Platforms (DAPs) have become essential tools for helping employees navigate complex enterprise software such as CRM, ERP, or HRMS systems. Companies like LemonLearning have shown how digital guidance can reduce training costs and accelerate onboarding. However, building and maintaining these interactive guides still requires extensive manual effort. Leveraging Large Language Models as virtual assistants is an appealing alternative, yet without a structured understanding of the target software, LLMs often hallucinate and produce unreliable answers. Moreover, most production-grade LLMs are black-box APIs, making fine-tuning impractical due to the lack of access to model weights. In this work, we introduce a Graph-based Retrieval-Augmented Generation framework that automatically converts enterprise web applications into state-action knowledge graphs, enabling LLMs to generate grounded and context-aware assistance. The framework was co-developed with the AI enterprise RAKAM, in collaboration with Lemon Learning. We detail the engineering pipeline that extracts and structures software interfaces, the design of the graph-based retrieval process, and the integration of our approach into production DAP workflows. Finally, we discuss scalability, robustness, and deployment lessons learned from industrial use cases.
Science Journals
arXiv:2604.04539v2 Announce Type: replace Abstract: Reinforcement learning (RL) is a core approach for robot control when expert demonstrations are unavailable. On-policy methods such as Proximal Policy Optimization (PPO) are widely used for their stability, but their reliance on narrowly distributed on-policy data limits accurate policy evaluation in high-dimensional state and action spaces. Off-policy methods can overcome this limitation by learning from a broader state-action distribution, yet suffer from slow convergence and instability, as fitting a value function over diverse data requires many gradient updates, causing critic errors to accumulate through bootstrapping. We present FlashSAC, a fast and stable off-policy RL algorithm built on Soft Actor-Critic. Motivated by scaling laws observed in supervised learning, FlashSAC sharply reduces gradient updates while compensating with larger models and higher data throughput. To maintain stability at increased scale, FlashSAC explicitly bounds weight, feature, and gradient norms, curbing critic error accumulation. Across over 60 tasks in 10 simulators, FlashSAC consistently outperforms PPO and strong off-policy baselines in both final performance and training efficiency, with the largest gains on high-dimensional tasks such as dexterous manipulation. In sim-to-real humanoid locomotion, FlashSAC reduces training time from hours to minutes, demonstrating the promise of off-policy RL for sim-to-real transfer.
arXiv:2605.16193v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used to simulate human opinions and survey responses, but their ability to reproduce population responses across cultures remains limited. Existing persona-based prompting methods typically rely on sociodemographic or personality traits, which are only indirect proxies for the values that shape human responses. We propose a value-based persona construction method that derives textual descriptors from survey responses capturing core cultural dimensions. By sampling value profiles from target populations and aggregating LLM responses across personas, we obtain population-level predictions grounded in observed value distributions. We further introduce a calibration procedure that improves response diversity while preserving estimated opinions. We show that our approach reduces prediction error across countries, with the largest improvements observed in underrepresented populations. This substantially narrows the performance gap between countries aligned with dominant LLM priors and those that are less represented in training data, while also yielding response distributions that closely match human diversity.
arXiv:2605.16232v1 Announce Type: new Abstract: The accelerating convergence of smart metering, generative artificial intelligence, and quantum-inspired combinatorial optimisation is reshaping how energy utilities manage physical infrastructure, customer engagement, and environmental accountability
arXiv:2605.15410v1 Announce Type: cross Abstract: Adaptive Non-local Observables (ANOs) have shown that making quantum observables dynamic can substantially enlarge the function space of Variational Quantum Algorithms, partly shifting hardware demands from circuit synthesis to measurement design. However, this advantage is accompanied by a steep increase in the number of parameters, as well as the classical optimization cost for varying general Hermitian observables. We propose a special form of ANO that significantly reduces this burden by considering only diagonal observables paired with quantum circuits. Mathematically, this is equivalent to the full ANO of a large parameter space since diagonal matrices are canonical representatives of the ANO space modulo unitary similarity. As a result, Diagonal ANO retains the same capability of full ANO while reducing $k$-local observable complexity from $O(4^k)$ to $O(2^k)$ and lowering the corresponding measurement-side classical computation. In this sense, diagonal ANO preserves much of the benefit of full ANO while encompassing conventional VQCs as a special case.
arXiv:2605.15927v1 Announce Type: new Abstract: Quantum chemistry calculations are often performed using atom-centered basis sets which are chosen to balance accuracy and cost. While they are systematically improvable, the total energy converges slowly with basis set size towards the complete basis set (CBS) limit. Common extrapolation methods require several intermediate-quality calculations to afford an estimate of the CBS energy. We propose combining a pairwise interaction model with a minimal complementary auxiliary basis set (CABS) baseline to estimate the CBS energy from a single quantum chemistry calculation in a minimal basis set via Kernel-Ridge-Regression (KRR), which is more efficient than both direct and $\Delta$-machine learning. We show that KRR on standard molecular representations can be improved by approximating atom-wise local kernels using Chebyshev polynomials which allows us to train KRR models efficiently on moderate compute resources, further enabling a data-driven approach towards CBS combining physical baselines capturing leading order effects with data-efficient machine learning models.
arXiv:2504.08300v5 Announce Type: replace Abstract: Benchmark-based evaluation, e.g., multiple-choice questions (MCQs) and open-ended questions (OEQs), is widely used for evaluating Large Language Models (LLMs), yet their reliability is undermined by benchmark contamination. When pre-exposed to the testing benchmark during training, less capable LLMs have been found to achieve inflated performance, thereby yielding erroneous results in LLM evaluation. In this study, we reframe contamination as an inherent aspect of learning and seek to disentangle and expose genuine capability acquisition from superficial memorization in LLM evaluation. Following this, firstly, by analyzing model performance under different memorization conditions of MCQs, we uncover a counterintuitive trend: LLMs perform worse on memorized benchmarks than on non-memorized ones, indicating the coexistence of two learning phenomena, i.e., rote memorization and genuine capability learning. To disentangle them, we propose TrinEval, a novel evaluation framework that reformulates MCQs into an alternative knowledge-centric trinity format, reducing memorization while preserving inherent knowledge, enabling the evaluation of genuine capability in the presence of memorization. Extensive experiments validate the effectiveness and robustness of TrinEval in reformulating benchmarks, and the evaluation results further reveal that mainstream LLMs rely on rote memorization for an average of 19.6% of knowledge points across the MMLU and the GSM8K dataset.
arXiv:2605.15805v1 Announce Type: cross Abstract: In typical single-molecule force spectroscopy experiments the mechanical unfolding of molecular complexes or biomolecules is studied applying a force ramp to one end of the system while the other end is kept fixed in space. The computational counterpart of this type of experiments can routinely be performed using molecular dynamics simulations with atomistic resolution. However, due to the large difference in time scales often coarse graining procedures are applied in the simulations. Most of the applied techniques do not allow to follow the atomistic details of the relevant conformational transitions due to the structural simplifications used to speed up the simulations. Here, we apply an earlier developed dynamic coarse graining technique based on Markov state modeling to a model peptidic system that does not unfold in a simple two-state manner. Using the donor-acceptor distances of the helical hydrogen bonds as collective variables and performing a dimension reduction technique allows us to construct a Markov model of the unfolding process that correctly represents the microscopic behavior of the system. The chosen example shows that the method can be used to mimick the mechanical unfolding process of systems for which the end-to-end distance does not provide a sufficient order parameter and that do not unfold in a simple cooperative manner.
arXiv:2605.15520v1 Announce Type: new Abstract: Data attribution has become an important component of pricing, auditing, and governance in machine learning pipelines, yet most attribution methods implicitly assume that attribution values faithfully reflect participants' contributions. We show that this assumption can fail: a single participant in a standard distributed training workflow can substantially inflate its measured attribution value while preserving global utility. Our attribution-first attack uses latent optimization to inject small synthetic batches that preserve utility while exploiting non-IID label coverage and evaluator sensitivities. Across datasets, models, and multiple marginal-utility evaluators, the attack consistently increases the adversary's attribution value and reshapes the relative attribution structure among benign clients without degrading accuracy or triggering geometry-based defenses. These results show that attribution itself forms a new attack surface and motivate the development of attribution-robust and incentive-compatible scoring mechanisms.
arXiv:2605.15956v1 Announce Type: new Abstract: Here we present a massive longitudinal dataset of public Telegram content, comprising over 5.9 billion messages dating from 2015 to 2025, collected from 712 thousand channels and groups, enriched with metadata on forwards, reactions, and polls. The dataset spans multiple languages including Russian and Farsi, representing countries where Telegram shows mainstream adoption, as well as Western languages where Telegram is used in specific sub-communities. The dataset has several advantages. First, when restricted by language, it provides a versatile example of an algorithm-free platform, contrary to many other social media platforms that are strongly influenced by opaque content-curation algorithms. Second, it enables comparative studies across different languages, communities, and user bases under identical platform affordances. The dataset thus offers a foundation for studying engagement patterns, network evolution, and community formation in the absence of algorithmic curation.
arXiv:2605.15514v1 Announce Type: new Abstract: We identify intrinsic limitations of Rotary Positional Embeddings (RoPE) in Transformer-based long-context language models. Our theoretical analysis abstracts away from the specific content of the context and depends only on its length. We prove that as context length increases, RoPE-based attention becomes unpredictable and loses two properties that are central to its effectiveness. First, it loses its locality bias: RoPE is no more likely to favor nearer positions than substantially farther ones. Second, it loses consistency in token relevance: a key vector that receives a higher attention score than an alternative at one position may receive a lower score at another. In both cases, the probability of failure approaches 0.5, no better than random guessing. We further prove that the attention score can remain unchanged when a key token is moved to a different position, or even replaced by a different token, indicating a failure to distinguish positions or tokens. Adjusting the RoPE base trades off distinguishing positions against distinguishing tokens but cannot preserve both at the same time. Increasing the RoPE base hyperparameter, a common practice in today's long-context models, helps distinguish different tokens, but inevitably sacrifices the ability to distinguish positions. Our empirical analysis shows that multi-head, multi-layer architectures are insufficient to overcome these limitations. Our findings suggest that fundamentally new mechanisms for encoding position and token order may be needed in future Transformer long-context language models.
arXiv:2412.06853v4 Announce Type: replace Abstract: This paper proposes a novel loss function, called 'Tube Loss', for simultaneous estimation of bounds of a Prediction Interval (PI) in the regression setup. The PIs obtained by minimizing the empirical risk based on the Tube Loss are shown to be of better quality than the PIs obtained by the existing methods in the following sense. First, it yields intervals that attain the prespecified confidence level t $\in$ (0,1) asymptotically. A theoretical proof of this fact is given. Secondly, the user is allowed to move the interval up or down by controlling the value of a parameter. This helps the user to choose a PI capturing denser regions of the probability distribution of the response variable inside the interval, and thus, sharpening its width. This is shown to be especially useful when the conditional distribution of the response variable is skewed. Further, the Tube Loss based PI estimation method can trade-off between the coverage and the average width by solving a single optimization problem. It enables further reduction of the average width of PI through re-calibration. Also, unlike a few existing PI estimation methods the gradient descent (GD) method can be used for minimization of empirical risk. Through extensive experiments, we demonstrate the effectiveness of Tube Loss-based PI estimation in both kernel machines and neural networks. Additionally, we show that Tube Loss-based deep probabilistic forecasting models achieve superior performance compared to existing probabilistic forecasting techniques across several benchmark and wind datasets. Finally, we empirically validate the advantages of the Tube loss approach within the conformal prediction framework. Codes are available at https://github.com/ltpritamanand/Tube$\_$loss.
arXiv:2605.16120v1 Announce Type: new Abstract: The growth of online video platforms drives the need for effective, semantically grounded event retrieval. We present MERVIN, a unified multimodal framework for Vietnamese news videos that integrates keyframes, transcripts, and video summaries. Transcript quality is enhanced via Gemini 1.5 Flash, reducing noise from accents, background sounds, and recognition errors. Visual features are extracted with Perception Encoder, while a Vietnamese language model produces textual embeddings; both are indexed in Milvus for efficient similarity-based retrieval. In addition, a React-based interface enables iterative query refinement across modalities, improving semantic alignment. Experimental results on Vietnamese news videos demonstrate the effectiveness of the proposed system, with MERVIN achieving 79 out of 88 points in AI Challenge HCMC 2025 qualification phase and successfully retrieved all results for every query in the final round.
arXiv:2602.18801v2 Announce Type: replace Abstract: Autoregressive neural PDE surrogates predict future states by repeatedly applying a learned one-step operator. This is a simple and widely used method, but small one-step errors can accumulate during long rollouts. The resulting drift often appears as spectral amplitude distortion, phase misalignment, and nonlinear mode-interaction error. These effects are especially important for time-dependent PDEs with clear Fourier structure. We introduce the Spectral Generator Neural Operator (SGNO), a structured autoregressive neural operator for long-horizon PDE forecasting. SGNO organizes each learned one-step map as a structured spectral evolution update. A real-valued nonpositive diagonal generator provides a gain-controlled spectral backbone, while a learned correction pathway with complex-valued spectral mixing completes the residual evolution. This design gives the autoregressive step an evolution-like structure while retaining the flexibility needed for dissipative, dispersive, transport-dominated, and nonlinear PDEs. SGNO is designed for periodic linear and semilinear evolution PDEs with Fourier multiplier linear dynamics. Across ten mechanism-matched APEBench tasks spanning this regime, SGNO consistently outperforms strong single-step autoregressive baselines in long-horizon rollout accuracy, reducing GMean100 by a median of 74.8% relative to the strongest available non-SGNO baseline, with per-task reductions ranging from 13.6% to 92.9%. The gains are strongest on dispersive and transport-dominated tasks, as well as tasks involving nonlinear closure and mode coupling. Spectral diagnostics show lower spectral energy error and improved rollout-level phase fidelity. Ablations show that the constrained generator, the structured update, and the learned correction pathway each contribute to performance. The code is available at https://github.com/cruiseresearchgroup/SGNO.
arXiv:2605.15881v1 Announce Type: cross Abstract: The modeling and simulation of infinite-dimensional Hamiltonian systems are central problems in mathematical physics and engineering, however they pose significant computational and structural challenges for standard data-driven architectures. In this work, we introduce the Symplectic Neural Operator, a neural operator architecture designed to preserve the symplectic structure intrinsic to Hamiltonian PDEs. We provide a theoretical characterization of their symplecticity and establish a rigorous long-term stability result based on the combination of symplectic structure preservation and learning accuracy. Numerical experiments on canonical Hamiltonian PDEs corroborate this theoretical result and show that SNOs exhibit improved energy behavior compared with non-structure-preserving neural operators.
arXiv:2602.16274v2 Announce Type: replace Abstract: We present the first regret bound for classical online Q-learning in infinite-horizon discounted Markov decision processes (MDPs), without relying on optimism or bonus terms. We first analyze Boltzmann Q-learning with decaying temperature and show that its regret depends critically on the suboptimality gap of the MDP: for sufficiently large gaps, the regret is sublinear, while for small gaps it deteriorates and can approach linear growth. To address this limitation, we study a Smoothed $\epsilon_n$-Greedy exploration scheme that combines $\epsilon_n$-greedy and Boltzmann exploration, for which we prove a gap-robust regret bound of near-$\tilde{O}(N^{9/10})$. We also obtain sample complexity guarantees, with both regret and sample complexity bounds holding with high probability. To analyze these algorithms, we develop a high-probability concentration bound for contractive Markovian stochastic approximation with iterate- and time-dependent transition dynamics. This bound may be of independent interest as the contraction factor in our framework is allowed to converge to one asymptotically.
arXiv:2602.09297v3 Announce Type: replace Abstract: Transformers update token representations through multi-head attention and residual connections as $X \leftarrow X + \sum_{i} P^{(i)}XW_{V_i}W_{o_i}$, where $P^{(i)}$ is the softmax attention matrix in head $i$. We propose replacing a subset of $P^{(i)}$'s with the Laplacian $I - P^{(i)}$, giving $X \leftarrow X + \sum_{i \in \mathcal{A}} P^{(i)}XW_{V_i}W_{o_i} + \sum_{i \in \mathcal{L}} (I - P^{(i)})XW_{V_i}W_{o_i}$. Our proposal has two motivations. First, it allows attention heads to update the mean of token representations, while Laplacian heads can directly control within-sequence variance. Second, if tokens are viewed as nodes in a graph with edge weights $P^{(i)}$, then $I - P^{(i)}$ is the corresponding graph Laplacian, and the update can be interpreted as one step of heat diffusion on the graph. We show that this simple modification improves performance across supervised learning, language modeling, and self-supervised learning tasks. To investigate why, we examine the token representations learned with and without Laplacian heads. In supervised learning, Laplacian heads collapse token representations within the same sequence and align the sequence means with the geometry of Neural Collapse. In language modeling, they increase the separability of token representations that share the same next-token prediction. In self-supervised learning, they produce token representations whose principal components are better suited for segmentation. Across modalities, they also lead to faster-decaying spectra, indicating stronger token smoothing. Overall, our findings challenge the prevailing view that token oversmoothing is inherently harmful, showing instead that certain forms of smoothing can be beneficial.
arXiv:2601.07820v2 Announce Type: replace Abstract: In human conversation, both interlocutors play an active role in maintaining mutual understanding. When listeners are uncertain about what speakers mean, for example, they can request clarification. It is an open question for language models whether they can assume a similar listener role, recognizing and expressing their own uncertainty through clarification. We argue that reference games are a suitable testbed to approach this question as they are controlled, self-contained, and make clarification needs explicit and measurable. To test this, we evaluate three vision-language models comparing a baseline reference resolution task to an experiment where the models are instructed to request clarification when uncertain. The results suggest that even in such simple tasks, models often struggle to recognize internal uncertainty and translate it into adequate clarification behavior. This demonstrates the value of reference games as testbeds for interaction qualities of (vision and) language models.
arXiv:2512.15693v2 Announce Type: replace Abstract: The misuse of AI-driven video generation technologies has raised serious social concerns, highlighting the urgent need for reliable AI-generated video detectors. However, most existing methods are limited to binary classification and lack the necessary explanations for human interpretation. In this paper, we present Skyra, a specialized multimodal large language model (MLLM) that identifies human-perceivable visual artifacts in AI-generated videos and leverages them as grounded evidence for both detection and explanation. To support this objective, we construct ViF-CoT-4K for Supervised Fine-Tuning (SFT), which represents the first large-scale AI-generated video artifact dataset with fine-grained human annotations. We then develop a two-stage training strategy that systematically enhances our model's spatio-temporal artifact perception, explanation capability, and detection accuracy. To comprehensively evaluate Skyra, we introduce ViF-Bench, a benchmark comprising 3K high-quality samples generated by over ten state-of-the-art video generators. Extensive experiments demonstrate that Skyra surpasses existing methods across multiple benchmarks, while our evaluation yields valuable insights for advancing explainable AI-generated video detection.
arXiv:2605.15533v1 Announce Type: new Abstract: Video editing poses a significant challenge. While a series of tuning-free methods circumvent the need for extensive data collection and model training, they often underutilize the rich information embedded within noisy latent, leading to unsatisfactory results. To address this, we propose a \textit{tuning-free, instruction-based} video editing framework. We approach video editing from the perspective of noisy latent: we design a Structural Noise Initialization Strategy (SNIS) to secure a superior editing starting point by assigning higher noise levels to edited regions (to facilitate content change) and lower noise levels to unedited regions (to maintain content consistency). We introduce a Noise Guidance Mechanism (NGM), which leverages the video prior in the generative model and effectively integrates rich information within the noisy latent to guide the denoising process, thereby preserving unedited content and overall visual coherence. Experiments show that our proposed method achieves better visual quality and state-of-the-art performance.
arXiv:2507.00275v2 Announce Type: replace Abstract: Double Q-learning is a classical control algorithm that mitigates the maximization bias of Q-learning. To do so, it explicitly trains two independent action-value functions and uses them to decouple action-selection and action-evaluation when computing bootstrap targets. Double DQN adapts target bootstrap decoupling to deep reinforcement learning (RL), but explicitly trains only a single action-value function and does not fully decouple its estimators. Consequently, the two estimators remain correlated, and overestimation persists. In this paper, we introduce Deep Double Q-learning (DDQL), a deep RL algorithm that explicitly trains two Q-functions through Double Q-learning. DDQL stabilizes training through a combination of techniques, including lower replay ratios, longer target network update intervals, and shared layers. Across 57 Atari 2600 games, DDQL improves aggregate performance over Double DQN, outperforming it on 47 games while further reducing overestimation. In addition, we study key design choices when adapting Double Q-learning to deep RL, including the network architecture, replay ratio, and minibatch sampling strategies.
arXiv:2506.23552v2 Announce Type: replace Abstract: The intrinsic link between facial motion and speech is often overlooked in generative modeling, where talking head synthesis and text-to-speech (TTS) are typically addressed as separate tasks. This paper introduces JAM-Flow, a unified framework to simultaneously synthesize and condition on both facial motion and speech. Our approach leverages flow matching and a novel Multi-Modal Diffusion Transformer (MM-DiT) architecture, integrating specialized Motion-DiT and Audio-DiT modules. These are coupled via selective joint attention layers and incorporate key architectural choices, such as temporally aligned positional embeddings and localized joint attention masking, to enable effective cross-modal interaction while preserving modality-specific strengths. Trained with an inpainting-style objective, JAM-Flow supports a wide array of conditioning inputs-including text, reference audio, and reference motion-facilitating tasks such as synchronized talking head generation from text, audio-driven animation, and much more, within a single, coherent model. JAM-Flow significantly advances multi-modal generative modeling by providing a practical solution for holistic audio-visual synthesis. project page: https://joonghyuk.com/jamflow-web
arXiv:2605.15715v1 Announce Type: new Abstract: Blockchain systems such as Solana or Monad employ tree- or star-shaped broadcast topologies in which a single source node disseminates message shards to a set of target peers within a strictly bounded time window. In these architectures, shard propagation must complete before the next consensus step, making timely delivery to a large fraction of the validator set essential. A fundamental limitation of such designs is that the outbound bandwidth of the source node constitutes the primary system bottleneck. In this paper, we introduce peer Turbo, a technique that allows target nodes to exchange shards using Random Linear Network Coding (RLNC), thereby assisting each other in completing decoding without requiring explicit shard state coordination. We use a tractable fluid approximation of the degree of freedom distribution of peer-Turbo-enabled systems show that this approach reduces source bandwidth required for a set service quality by up to one order of magnitude, or equivalently reduces propagation latency by one order of magnitude under fixed bandwidth constraints.
arXiv:2605.15450v1 Announce Type: new Abstract: Concealed Object Segmentation (COS) encompasses a family of dense-prediction tasks, including camouflaged object detection, polyp segmentation, transparent object detection, and industrial defect inspection, where targets are visually entangled with their surroundings through different physical mechanisms. Existing methods either operate directly on RGB images or employ \emph{heterogeneous} decompositions (\eg, Fourier, wavelet) that redistribute spatial evidence across scale/frequency coefficients, making pixel-aligned cues less direct. We introduce a fundamentally different perspective: \textbf{homogeneous image decomposition} via Retinex theory, which factorizes an image into illumination and reflectance components within the \emph{same} spatial domain. Our key insight is that visual entanglement enforces appearance matching in the composite space, but this does \emph{not} necessitate simultaneous matching in both component spaces, a phenomenon we formalize as the \textbf{Discriminability Gap Theorem}. Crucially, we show that across diverse COS sub-tasks, the underlying physical processes systematically anti-correlate illumination and reflectance differences, yielding theoretical guarantees that Retinex decomposition preserves or strictly improves total foreground--background discriminability across the full physical regime, with anti-correlation maximizing the gain. Building on this, we propose \textbf{RIDE} comprising: (i) a Task-Driven Retinex Decomposition module that learns segmentation-optimal factorizations end-to-end; (ii) a Discriminability Gap Attention mechanism that adaptively exploits where decomposition helps; and (iii) a Camouflage-Breaking Contrastive loss operating in reflectance feature space.
arXiv:2605.16113v1 Announce Type: new Abstract: Large language models (LLMs) have achieved unprecedented success due to their exceptional generative capabilities. However, because they depend on knowledge encapsulated from training corpora, they may produce hallucinations, stereotypes, and socially biased content. In particular, LLMs are prone to prejudiced responses involving race, gender, and age, which are collectively referred to as social biases. Prior studies have used fine-tuning and prompt engineering to mitigate such biases in LLMs, but these methods require additional training resources or domain knowledge to design the framework. Moreover, they may degrade the original capabilities of LLMs and often overlook the need for dynamic debiasing contexts for fairer inference. In this paper, we propose DebiasRAG, a novel tuning-free and dynamic query-specific debiasing framework based on retrieval-augmented generation (RAG). DebiasRAG improves fairness while preserving the intrinsic properties of LLMs, such as representation ability. DebiasRAG consists of three stages: (1) query-specific debiasing candidate generation; (2) context candidate pool construction; and (3) gradient-updated debiasing-guided context piece reranking. First, DebiasRAG leverages self-diagnosed bias contexts relevant to the query through regular retrieval, where the bias contexts are prepared offline by the DebiasRAG provider. Given the query-specific bias contexts, DebiasRAG reversely produces debiasing contexts, which are provided as additional fairness constraints for LLM outputs. Second, a regular RAG retrieval process produces query-related contexts from the regular RAG document database, such as a chunked Wikipedia dataset.