arXiv subjects

Santu Rana

Publications and source records attributed to Santu Rana.

At least 19 recordsLinked to original sources

arXiv2026

DISEIL: Demonstration Distillation for Sample-Efficient Imitation Learning

A robot that can be taught a new task from a handful of demonstrations has to work out for itself what it still cannot do, and then ask for exactly that. Interactive imitation learning takes a step in that direction by letting a policy practice on its own and calling an expert when it goes wrong. Existing methods decide when to interrupt the learner. A further 2 decisions are left to whichever episode happened to trigger the interruption: which failure to correct, and where the demonstration should start. This paper is a first attempt at making both of them deliberately. DISEIL (Demonstration dIstillation for Sample-Efficient Imitation Learning) marks each failed episode at the step where the policy first becomes unreliable, represents that moment with a geometric descriptor, and groups the failures into recurring failure modes. A vision-language model and a language model read the selected mode and write a request for the next demonstration, and a store of task constraints checks that the request can be carried out before any expert time is spent. No model produces a robot action. Across 5 simulated tasks under state and image observations, changing only what the expert is asked for gives the highest mean held-out success rate in all 10 settings, with a tie in 1, and the margin is widest at the smallest budget we tested. The scope is narrow: a single round of practice at a time, in simulation, with experts that are mostly scripted. The longer-term aim is a learner that also tracks what its demonstration set already covers, and that asks a human teacher for the missing behavior in proportion to the effort each request costs them.

cs.RO↗

arXiv2026

MAGIK: Mapping to Analogous Goals via Imagination-enabled Knowledge Transfer

Humans excel at analogical reasoning - applying knowledge from one task to a related one with minimal relearning. In contrast, reinforcement learning (RL) agents typically require extensive retraining even when new tasks share structural similarities with previously learned ones. In this work, we propose MAGIK, a novel framework that enables RL agents to transfer knowledge to analogous tasks without interacting with the target environment. Our approach leverages an imagination mechanism to map entities in the target task to their analogues in the source domain, allowing the agent to reuse its original policy. Experiments on custom MiniGrid and MuJoCo tasks show that MAGIK achieves effective zero-shot transfer using only a small number of human-labelled examples. We compare our approach to related baselines and highlight how it offers a novel and effective mechanism for knowledge transfer via imagination-based analogy mapping.

cs.AI↗

arXiv2026

Hidden in Thought: Transferable Chain-of-Thought Artifacts Induce Harmful Behavior

We investigate whether harmful chain-of-thought (CoT) traces from compromised language models can transfer unsafe behaviour and be distilled into reusable jailbreak attacks. Using an emergent-misalignment organism and a refusal-ablated jailbroken organism, we transplant harmful CoTs into $29$ open-source and $5$ closed-source targets. Transferred traces raise harmful-response rates above $80\%$ on the most vulnerable open-source models, while semantically mismatched CoTs fail entirely. LLooM concept mining identifies four recurring components of harmful reasoning: proceduralisation, ethical decoupling, evasion, and target--vulnerability framing. Distilling these patterns into reusable system prompts produces effective black-box jailbreaks, outperforming direct CoT transplantation on strongly aligned models by up to an order of magnitude, including a $10\times$ improvement on GPT-4.1 AdvBench. Reasoning-enabled models are more than twice as vulnerable, and output-side safeguards such as Llama-Guard~3 frequently miss harmful generations. Our results show that harmful reasoning transfers at both the trace and pattern levels, motivating defences that evaluate reasoning context in addition to final outputs.

cs.CR↗

arXiv2026

OnlyDense: Reduced-Order Modeling for Lagrangian simulation

In science and engineering, Lagrangian simulation methods such as Smooth Particle Hydrodynamics (SPH) or Material Point Method (MPM) are often employed to study the behavior of dynamic systems. However, these methods can be prohibitively computationally expensive, particularly when simulating multi-scale spatial or temporal phenomena, e.g., void growth and coalescence within macro-scale geometries, structural failure of spacecraft components resulting from hypervelocity impact of space debris particles, etc. In contrast to graph-based methods, where the state of the system is understood as a discrete set of particles, we propose a learning framework for scalable representation and dynamics modeling of massive particle systems by treating the system state as a function and its evolution as a trajectory in Hilbert space. Rather than representing the state as a discrete set of particles or embedding it in a nonlinear latent manifold, we approximate the state space with a linear subspace spanned by learned neural basis functions. This parameterization enables direct projection to obtain latent coefficients and explicit access to the basis functions, avoiding optimization over a nonlinear latent space. The resulting representation admits a natural interpretation: latent variables correspond to coefficients in Hilbert space, and basis functions correspond to spatial modes, analogous to Proper Orthogonal Decomposition. The framework thus unifies classical projection-based reduced-order modeling with modern deep learning, while remaining invariant to the number of discretization points. Experiments on large-scale SPH simulations with over one million particles, including dynamic events with extreme deformation and fragmentation, demonstrate that the proposed method accurately reconstructs and predicts dynamics, achieving an R$^2$ score above $0.99$ with as few as $32$ basis functions.

cs.LG↗

arXiv2026

Shared Latent Structures Enable Unified Backdoor Detection and Mitigation in LLMs

Backdoor attacks in large language models (LLMs) are often treated as isolated trigger-response failures, motivating defenses tailored to specific triggers or behaviors. We show this view is incomplete. Across diverse backdoor behaviors, we identify a shared latent mechanism that can be detected, causally controlled, and suppressed. Using sparse autoencoders (SAEs) on residual-stream activations, we find a small set of latent features consistently activated across jailbreaking, refusal manipulation, password-locking, bias induction, sentiment misclassification, and country-conditioned harmful advice. These features generalize across Qwen3, Gemma~3, and Llama~3.1 models from 4B to 32B parameters, and across both fine-tuning and weight-editing attacks. Through bidirectional activation steering, we show these features are causal: suppressing them reduces attack success, while amplifying them induces target behaviors on clean prompts. We further train lightweight SAE-feature classifiers that generalize zero-shot to unseen backdoors and outperform residual-stream and weight-diffing baselines. Finally, we introduce Concept Ablation Fine-Tuning (CAFT), which suppresses backdoor formation by ablating the shared latent subspace during training. Together, our results suggest that many backdoors rely on a transferable latent mechanism, enabling unified detection and mitigation.

cs.AI↗

arXiv2026

Leveraging Human Feedback for Semantically-Relevant Skill Discovery

Unsupervised skill discovery in reinforcement learning aims to intrinsically motivate agents to discover diverse and useful behaviours. However, unconstrained approaches can produce unsafe, unethical, or misaligned behaviours. To mitigate these risks and improve the practical desireability of discovered skills, recent work grounds the discovery process by leveraging human preference feedback. However, preference-based approaches are feedback-inefficient and inherently ill-equipped to deal with skill spaces composed of a variety of different skills such as running, jumping, walking, etc. To overcome this limitation, we introduce semantic labelling, a novel and feedback-efficient approach that leverages human cognitive strengths to identify and label semantically meaningful behaviours. Based on semantic labelling, we propose Semantically Relevant Skill Discovery (SRSD), a novel human-in-the-loop approach that collects semantic labels from human feedback and learns a reward function to encourage skills to be more semantically diverse and relevant. Through our experiments in a 2D navigation environment and four locomotion environments, we demonstrate that SRSD can improve semantic diversity and discover relevant behaviours while scaling effectively to a large variety of behaviours.

cs.LG↗

arXiv2026

ASPECT:Analogical Semantic Policy Execution via Language Conditioned Transfer

Reinforcement Learning (RL) agents often struggle to generalize knowledge to new tasks, even those structurally similar to ones they have mastered. Although recent approaches have attempted to mitigate this issue via zero-shot transfer, they are often constrained by predefined, discrete class systems, limiting their adaptability to novel or compositional task variations. We propose a significantly more generalized approach, replacing discrete latent variables with natural language conditioning via a text-conditioned Variational Autoencoder (VAE). Our core innovation utilizes a Large Language Model (LLM) as a dynamic \textit{semantic operator} at test time. Rather than relying on rigid rules, our agent queries the LLM to semantically remap the description of the current observation to align with the source task. This source-aligned caption conditions the VAE to generate an imagined state compatible with the agent's original training, enabling direct policy reuse. By harnessing the flexible reasoning capabilities of LLMs, our approach achieves zero-shot transfer across a broad spectrum of complex and truly novel analogous tasks, moving beyond the limitations of fixed category mappings. Code and videos are available \href{https://anonymous.4open.science/r/ASPECT-85C3/}{here}.

cs.AI↗

arXiv2026

The Unintended Trade-off of AI Alignment:Balancing Hallucination Mitigation and Safety in LLMs

Hallucination in large language models (LLMs) has been widely studied in recent years, with progress in both detection and mitigation aimed at improving truthfulness. Yet, a critical side effect remains largely overlooked: enhancing truthfulness can negatively impact safety alignment. In this paper, we investigate this trade-off and show that increasing factual accuracy often comes at the cost of weakened refusal behavior. Our analysis reveals that this arises from overlapping components in the model that simultaneously encode hallucination and refusal information, leading alignment methods to suppress factual knowledge unintentionally. We further examine how fine-tuning on benign datasets, even when curated for safety, can degrade alignment for the same reason. To address this, we propose a method that disentangles refusal-related features from hallucination features using sparse autoencoders, and preserves refusal behavior during fine-tuning through subspace orthogonalization. This approach prevents hallucinations from increasing while maintaining safety alignment.We evaluate our method on commonsense reasoning tasks and harmful benchmarks (AdvBench and StrongReject). Results demonstrate that our approach preserves refusal behavior and task utility, mitigating the trade-off between truthfulness and safety.

cs.CL↗

arXiv2025

Dynamic Policy Fusion for User Alignment Without Re-Interaction

Deep reinforcement learning (RL) policies, although optimal in terms of task rewards, may not align with the personal preferences of human users. To ensure this alignment, a naive solution would be to retrain the agent using a reward function that encodes the user's specific preferences. However, such a reward function is typically not readily available, and as such, retraining the agent from scratch can be prohibitively expensive. We propose a more practical approach - to adapt the already trained policy to user-specific needs with the help of human feedback. To this end, we infer the user's intent through trajectory-level feedback and combine it with the trained task policy via a theoretically grounded dynamic policy fusion approach. As our approach collects human feedback on the very same trajectories used to learn the task policy, it does not require any additional interactions with the environment, making it a zero-shot approach. We empirically demonstrate in a number of environments that our proposed dynamic policy fusion approach consistently achieves the intended task while simultaneously adhering to user-specific needs.

cs.AI↗

arXiv2025

Improving Multilingual Language Models by Aligning Representations through Steering

This paper investigates how Large Language Models (LLMs) represent non-English tokens -- a question that remains underexplored despite recent progress. We propose a lightweight intervention method using representation steering, where a learned vector is added to the residual stream at a single model layer to enhance multilingual performance. Through extensive experiments across seven competitive baselines -- including prompt optimization, supervised fine-tuning (SFT), in-context learning, cross-lingual transfer, and translation-based methods-we show that our approach consistently outperforms most alternatives. In particular, it achieves performance on par with production-grade translation systems while requiring far fewer resources. We further explore the complementarity between our method and SFT, demonstrating that steering offers a direct, efficient way to realign internal representations. These findings underscore the potential of activation-level interventions as a powerful tool for improving the multilingual capabilities of LLMs.

cs.CL↗

arXiv2025

Learning coupled Allen-Cahn and Cahn-Hilliard phase-field equations using Physics-informed neural operator(PINO)

Phase-field equations, mostly solved numerically, are known for capturing the mesoscale microstructural evolution of a material. However, such numerical solvers are computationally expensive as it needs to generate fine mesh systems to solve the complex Partial Differential Equations(PDEs) with good accuracy. Therefore, we propose an alternative approach of predicting the microstructural evolution subjected to periodic boundary conditions using Physics informed Neural Operators (PINOs). In this study, we have demonstrated the capability of PINO to predict the growth of $θ^{\prime}$ precipitates in Al-Cu alloys by learning the operator as well as by solving three coupled physics equations simultaneously. The coupling is of two second-order Allen-Cahn equation and one fourth-order Cahn-Hilliard equation. We also found that using Fourier derivatives(pseudo-spectral method and Fourier extension) instead of Finite Difference Method improved the Cahn-Hilliard equation loss by twelve orders of magnitude. Moreover, since differentiation is equivalent to multiplication in the Fourier domain, unlike Physics informed Neural Networks(PINNs), we can easily compute the fourth derivative of Cahn-Hilliard equation without converting it to coupled second order derivative.

cs.CE↗

arXiv2025

TRUST: Test-time Resource Utilization for Superior Trustworthiness

Standard uncertainty estimation techniques, such as dropout, often struggle to clearly distinguish reliable predictions from unreliable ones. We attribute this limitation to noisy classifier weights, which, while not impairing overall class-level predictions, render finer-level statistics less informative. To address this, we propose a novel test-time optimization method that accounts for the impact of such noise to produce more reliable confidence estimates. This score defines a monotonic subset-selection function, where population accuracy consistently increases as samples with lower scores are removed, and it demonstrates superior performance in standard risk-based metrics such as AUSE and AURC. Additionally, our method effectively identifies discrepancies between training and test distributions, reliably differentiates in-distribution from out-of-distribution samples, and elucidates key differences between CNN and ViT classifiers across various vision datasets.

cs.LG↗

arXiv2025

Alpaca against Vicuna: Using LLMs to Uncover Memorization of LLMs

In this paper, we introduce a black-box prompt optimization method that uses an attacker LLM agent to uncover higher levels of memorization in a victim agent, compared to what is revealed by prompting the target model with the training data directly, which is the dominant approach of quantifying memorization in LLMs. We use an iterative rejection-sampling optimization process to find instruction-based prompts with two main characteristics: (1) minimal overlap with the training data to avoid presenting the solution directly to the model, and (2) maximal overlap between the victim model's output and the training data, aiming to induce the victim to spit out training data. We observe that our instruction-based prompts generate outputs with 23.7% higher overlap with training data compared to the baseline prefix-suffix measurements. Our findings show that (1) instruction-tuned models can expose pre-training data as much as their base-models, if not more so, (2) contexts other than the original training data can lead to leakage, and (3) using instructions proposed by other LLMs can open a new avenue of automated attacks that we should further study and explore. The code can be found at https://github.com/Alymostafa/Instruction_based_attack .

cs.CL↗

arXiv2025

Human-Aligned Skill Discovery: Balancing Behaviour Exploration and Alignment

Unsupervised skill discovery in Reinforcement Learning aims to mimic humans' ability to autonomously discover diverse behaviors. However, existing methods are often unconstrained, making it difficult to find useful skills, especially in complex environments, where discovered skills are frequently unsafe or impractical. We address this issue by proposing Human-aligned Skill Discovery (HaSD), a framework that incorporates human feedback to discover safer, more aligned skills. HaSD simultaneously optimises skill diversity and alignment with human values. This approach ensures that alignment is maintained throughout the skill discovery process, eliminating the inefficiencies associated with exploring unaligned skills. We demonstrate its effectiveness in both 2D navigation and SafetyGymnasium environments, showing that HaSD discovers diverse, human-aligned skills that are safe and useful for downstream tasks. Finally, we extend HaSD by learning a range of configurable skills with varying degrees of diversity alignment trade-offs that could be useful in practical scenarios.

cs.LG↗

arXiv2024

Efficient Symmetry-Aware Materials Generation via Hierarchical Generative Flow Networks

Discovering new solid-state materials requires rapidly exploring the vast space of crystal structures and locating stable regions. Generating stable materials with desired properties and compositions is extremely difficult as we search for very small isolated pockets in the exponentially many possibilities, considering elements from the periodic table and their 3D arrangements in crystal lattices. Materials discovery necessitates both optimized solution structures and diversity in the generated material structures. Existing methods struggle to explore large material spaces and generate diverse samples with desired properties and requirements. We propose the Symmetry-aware Hierarchical Architecture for Flow-based Traversal (SHAFT), a novel generative model employing a hierarchical exploration strategy to efficiently exploit the symmetry of the materials space to generate crystal structures given desired properties. In particular, our model decomposes the exponentially large materials space into a hierarchy of subspaces consisting of symmetric space groups, lattice parameters, and atoms. We demonstrate that SHAFT significantly outperforms state-of-the-art iterative generative methods, such as Generative Flow Networks (GFlowNets) and Crystal Diffusion Variational AutoEncoders (CDVAE), in crystal structure generation tasks, achieving higher validity, diversity, and stability of generated structures optimized for target properties and requirements.

cs.LG↗

arXiv2024

ECoDe: A Sample-Efficient Method for Co-Design of Robotic Agents

Co-designing autonomous robotic agents involves simultaneously optimizing the controller and physical design of the agent. Its inherent bi-level optimization formulation necessitates an outer loop design optimization driven by an inner loop control optimization. This can be challenging when the design space is large and each design evaluation involves a data-intensive reinforcement learning process for control optimization. To improve the sample efficiency of co-design, we propose a multi-fidelity-based exploration strategy in which we tie the controllers learned across the design spaces through a universal policy learner for warm-starting subsequent controller learning problems. Experiments performed on a wide range of agent design problems demonstrate the superiority of our method compared to baselines. Additionally, analysis of the optimized designs shows interesting design alterations, including design simplifications and non-intuitive alterations.

cs.RO↗

arXiv2024

A Data-Driven Defense against Edge-case Model Poisoning Attacks on Federated Learning

Federated Learning systems are increasingly subjected to a multitude of model poisoning attacks from clients. Among these, edge-case attacks that target a small fraction of the input space are nearly impossible to detect using existing defenses, leading to a high attack success rate. We propose an effective defense using an external defense dataset, which provides information about the attack target. The defense dataset contains a mix of poisoned and clean examples, with only a few known to be clean. The proposed method, DataDefense, uses this dataset to learn a poisoned data detector model which marks each example in the defense dataset as poisoned or clean. It also learns a client importance model that estimates the probability of a client update being malicious. The global model is then updated as a weighted average of the client models' updates. The poisoned data detector and the client importance model parameters are updated using an alternating minimization strategy over the Federated Learning rounds. Extensive experiments on standard attack scenarios demonstrate that DataDefense can defend against model poisoning attacks where other state-of-the-art defenses fail. In particular, DataDefense is able to reduce the attack success rate by at least ~ 40% on standard attack setups and by more than 80% on some setups. Furthermore, DataDefense requires very few defense examples (as few as five) to achieve a near-optimal reduction in attack success rate.

cs.LG↗

arXiv2024

Composite Concept Extraction through Backdooring

Learning composite concepts, such as \textquotedbl red car\textquotedbl , from individual examples -- like a white car representing the concept of \textquotedbl car\textquotedbl{} and a red strawberry representing the concept of \textquotedbl red\textquotedbl -- is inherently challenging. This paper introduces a novel method called Composite Concept Extractor (CoCE), which leverages techniques from traditional backdoor attacks to learn these composite concepts in a zero-shot setting, requiring only examples of individual concepts. By repurposing the trigger-based model backdooring mechanism, we create a strategic distortion in the manifold of the target object (e.g., \textquotedbl car\textquotedbl ) induced by example objects with the target property (e.g., \textquotedbl red\textquotedbl ) from objects \textquotedbl red strawberry\textquotedbl , ensuring the distortion selectively affects the target objects with the target property. Contrastive learning is then employed to further refine this distortion, and a method is formulated for detecting objects that are influenced by the distortion. Extensive experiments with in-depth analysis across different datasets demonstrate the utility and applicability of our proposed approach.

cs.CV↗