Machine Learning · Cross-listed from cs.CL
Learning When to Think While Listening in Large Audio-Language Models
Zhiyuan Song, Weici Zhao, Yang Xiao et al. (+3 more)
Recent advances in Large Audio-
Language Models (LALMs) have made real-time, streaming spoken interaction increasingly practical. In this setting,
reasoning quality and responsiveness are tightly coupled: delaying
reasoning until the
speech endpoint can improve answer quality but moves deliberation into user-visible response delay, while answering too early risks committing before decisive evidence arrives. We introduce a learnable wait-think-answer control formulation for LALMs. Motivated by the incremental nature of human conversation, the controller decides under partial audio evidence when to wait, when to externalize a compact
reasoning update, and when to answer. Using Qwen2.5-Omni-7B as the base model, we construct aligned wait-think-answer traces from spoken
reasoning data, train the controller with supervised fine-tuning (SFT), and then apply Decoupled Clip and Dynamic Sampling Policy Optimization (DAPO). The reward combines answer correctness, action validity, update timing, latency synchronization,
reasoning quality, and chain consistency, optimizing the complete wait-think-answer trajectory and not the final answer alone. On a six-task synthetic spoken
reasoning question answering (SRQA)
benchmark, the six-reward DAPO controller improves the row-weighted accuracy from 67.6% to 70.3% while reducing post-endpoint final-think length by 14% under the same Qwen deployment harness. On a 186-item human-recorded Real Audio Bench, a transfer check beyond text-to-
speech (TTS)-rendered
speech, the controller family remains functional: SFT achieves the strongest accuracy, while the six-reward DAPO controller is the only learned variant whose final-think length falls below the base. These results suggest that a streaming model should learn when to make intermediate
reasoning explicit during the audio stream.
Read the full abstract ▾
All authors Zhiyuan Song, Weici Zhao, Yang Xiao, Suhao Yu, Cheng Zhu, Jiatao Gu
Machine Learning · Cross-listed from cs.CL
Beyond Binary: Speech Representations Across the Cognitive Score Hierarchy
Serli Kopar, Roshan Prakash Rane, Christian Mychajliw et al. (+6 more)
This study examines the relationship between
speech representations and the hierarchical structure of cognitive assessment in mild cognitive impairment. Utilizing 5,754 German neuropsychological assessment recordings, we evaluate six cognitive tasks across three score levels: task, domain, and global levels. We compare hand-crafted acoustic features with self-supervised learning (SSL)
embeddings. Results show that although SSL representations generally outperform hand-crafted features at lower levels, this trend reverses for MCI classification. Furthermore, task-specific constraints influence performance: tasks with greater response freedom exhibit performance dilution as hierarchical levels increase, suggesting “specialist” representations, whereas the performance of highly structured tasks increases toward higher levels, suggesting “generalist” representations. These findings show links between task constraints and assessment hierarchy in automated clinical
speech analysis.
Read the full abstract ▾
All authors Serli Kopar, Roshan Prakash Rane, Christian Mychajliw, Lydia Federmann, Gerhard Eschweiler, Daniela Berg, Sam Gijsen, Paula Andrea Perez-Toro, Kerstin Ritter
Machine Learning · Cross-listed from cs.CV
FoundObj: Self-supervised Foundation Models as Rewards for Label-free 3D Object Segmentation
Zihui Zhang, Zhixuan Sun, Yafei Yang et al. (+3 more)
We address the challenging task of 3D object
segmentation in complex scene point clouds without relying on any scene-level human annotations during training. Existing methods are typically constrained to identifying simple objects, primarily due to insufficient object priors in the learning process. In this paper, we present FoundObj, a novel framework featuring a superpoint-based object discovery
agent that incrementally merges suitable neighboring superpoints, guided by our innovative semantic and
geometric reward modules. These modules synergistically leverage semantic and
geometric priors from self-supervised 2D/3D foundation models, providing complementary feedback to the object discovery
agent and enabling robust identification of multi-class objects through reinforcement learning. Extensive experiments on diverse
benchmarks demonstrate that our approach consistently outperforms existing baselines. Notably, our method exhibits strong generalization in zero-shot and long-tail scenarios, underscoring its potential for scalable, label-free 3D object
segmentation.
Read the full abstract ▾
All authors Zihui Zhang, Zhixuan Sun, Yafei Yang, Jinxi Li, Jiahao Chen, Bo Yang
Machine Learning
The Role of Causal Features in Strategic Classification for Robustness and Alignment
Antonio Gois, Sophia Gunluk, Nir Rosenfeld et al. (+3 more)
In strategic classification, an institution (e.g., a bank) anticipates adaptation from users who change their features to increase utility in a classification task (e.g., loan repayment). Since a key challenge is the distribution shift induced by users, we turn to causal models, which have been shown to bound the worst-case out-of-distribution (OOD) risk, and establish several new results that link causality and strategic classification. First, we show that causal classification leads to optimal classification error after any sufficiently large adaptation, when the noise is bounded in a certain way. Second, when these assumptions do not hold, we show OOD cross-
entropy risk of optimal classifiers decomposes into an OOD bias term and a term arising from not using all observable features, allowing us to understand when causal classifiers have an advantage. Finally, we show that the use of causal features can allow
alignment of long-term incentives between institutions and users, contrasting with previous work that highlights social costs of such approaches. We validate our theory empirically on synthetic data, finding that our results predict behavior in practice.
Read the full abstract ▾
All authors Antonio Gois, Sophia Gunluk, Nir Rosenfeld, Nidhi Hegde, Simon Lacoste-Julien, Dhanya Sridhar
Machine Learning · Cross-listed from cs.CV
Is an Image Also Worth 16x16=256 Superpixels? A Framework for Attentional Image Classification
Pedro Henrique da Costa Avelar, Anderson R. Tavares and Luís C. Lamb
Superpixel-based image classification has traditionally leveraged graph neural networks (GNNs) for processing irregular image representations. Recent advances in computer
vision, driven by
Vision Transformers (ViTs), have introduced new paradigms in self-attentional models, surpassing convolutional neural networks (CNNs) in various tasks. However, a synergistic connection between GNNs, superpixels, and
transformers remains unexplored. In this work, we propose Superpixel
Transformers (SPT), a novel framework that unifies superpixel-based image classification and ViTs. SPT generalizes the Superpixel Image Classification with Graph
Attention Networks (SICGAT) model and ViT to support arbitrary superpixel-based chunking strategies, connectivity graphs, and positional encodings. We introduce refinements including a multidimensional sine-cosine positional encoding and an enriched patch data structure that fully incorporates superpixel shape and color information. By testing SPT across
datasets such as CIFAR10, FashionMNIST, and Imagenette, with various superpixel generation and graph connectivity strategies, we demonstrate that SPT achieves superior performance compared to previous superpixel-based GNN methods and remains competitive with ViTs. Notably, our approach addresses the limitations of SICGAT, such as information loss during pixel aggregation, and shows how constrained graph connectivity can enhance ViT performance. SPT bridges the gap between superpixel-based and
transformer models, opening avenues for cross-domain generalization and future innovations in hybrid attentional frameworks, and showing that an image can also be worth $16\times16$ superpixels.
Read the full abstract ▾
Machine Learning
Deep-layer limit and stability analysis of the basic forward-backward-splitting induced network (II): learning problems
Xuan Lin and Chunlin Wu
Deep unfolding neural networks derived from iterative optimization schemes and numerical ordinary/partial differential equations (ODEs/PDEs) have attracted much
attention in data science over the last decade. Therein, numerous important network architectures were constructed from the basic forward-backward-splitting (FBS)
algorithm. In this paper, we continue our research on the most basic FBS-induced network, an architecture unrolled from the original FBS
algorithm by incorporating direct parameter relaxations. Following the difference/differential inclusion formulations in our previous forward system analyses, we here consider some theoretical aspects of corresponding learning problems. Under some mild assumptions, we establish a general convergence property of the training problem of the basic FBS-induced network to the learning problem of the deep-layer limit system, implying a $\Gamma$-convergence argument showing that any cluster point of the optimal learning parameters for the network is a solution to the learning problem of the deep-layer limit system. A qualitative analysis of perturbation stabilities of these learning problems is also presented. A simple numerical experiment is conducted to validate our main general convergence result.
Read the full abstract ▾
Machine Learning
DEI: Diversity in Evolutionary Inference for Quality-Diversity Search
John Donaghy and Shikhar Rastogi
We present DEI: Diversity in Evolutionary Inference, a distributed Quality-Diversity (QD) search framework that assigns heterogeneous large
language models (LLMs) as mutation operators across peer nodes communicating with non-blocking collective operations. Unlike homogeneous parallel search, which replicates a single model's inductive biases across all workers, DEI treats each LLM's distinct creative prior as a complementary source of behavioral novelty. Extending the Digital Red Queen framework with DEI, nodes share local optimal solutions at the end of each round to seed the next round's population. This creates cross-model
adversarial pressure that drives robustness beyond intra-model self-play. Evaluated on the Core War domain, a competitive programming
benchmark in which Redcode warrior programs battle inside a simulated machine, a four-node heterogeneous ensemble (GPT-5.4-mini, Claude Sonnet 4.6, GPT-5.2, and Claude Haiku 4.5) achieves 124 percent higher merged-archive QD-Score (45.90 vs. 20.46) and 28 percent higher coverage (80.6 percent vs. 63.0 percent of
cells) than a single-node baseline at equal total LLM-call budget. The heterogeneous ensemble also outperforms an equally-budgeted homogeneous ensemble on QD-Score, coverage, and held-out solution generality across all four model families. These results provide the first empirical evidence that model diversity, not merely parallelism, is the key driver of gain in distributed LLM-based QD search.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.CV
PILOT: A Data-Free Continual Learning Approach for Real-Time Semantic Segmentation via Boundary Guidance
Yujing Zhou, Prashant Shekhar, Thomas Yang et al. (+1 more)
Real-time semantic
segmentation models offer an excellent balance between accuracy and inference speed. However, deploying these models in dynamic real world environments often requires the ability to learn novel classes incrementally without retraining on the entire
dataset. This capability is known as continual learning. In this regard, the standard fine-tuning methods in deep learning often fail due to catastrophic forgetting, where the model learns new information but forgets previously trained and learned classes. Contributing to this crucial domain, the current paper proposes a novel continual learning framework tailored for PIDNet, which is a widely cited state-of-the-art real-time semantic
segmentation model. Our method, PILOT(Parallel Incremental Learning Over Time), introduces a real-time and lightweight strategy by implementing a parallel Derivative-branch (D-branch) designed to capture the high frequency boundary information of novel classes while freezing the trained parameters of the original
segmentation network. This novel setup allows the model to adapt to new semantic categories while preserving the knowledge of previously learned classes. By using only data associated with the new class, our model significantly reduces training overhead. Experimental results demonstrate that our approach successfully segments new classes while maintaining high mean Intersection over Union (mIoU) on the original base classes, thereby comfortably outperforming all major continual learning approaches in this domain. Overall, PILOT is shown to effectively mitigate catastrophic forgetting with minimal impact on inference latency, thus maintaining real-time performance.
Read the full abstract ▾
All authors Yujing Zhou, Prashant Shekhar, Thomas Yang, Yongxin Liu
Machine Learning
High-Quality Synthetic Financial Time-Series using a GAN-Diffusion Framework
Giuseppe Masi, Andrea Coletta and Novella Bartolini
In recent years, financial institutions and firms have increasingly adopted synthetic data to address data scarcity and to generate counterfactual
market scenarios. However, reproducing all the statistical properties of financial time series, commonly known as stylized facts, remains an open challenge for many existing general-purpose architectures. In this paper, we present a quality-aware generative framework that combines two classes of generative methods, demonstrating how their integration addresses existing limitations while enhancing the realism of synthetic data. Specifically, we first introduce CoMeTS-GAN (Correlated Multivariate Time Series GAN), a Conditional Generative
Adversarial Network (C-GAN) designed to jointly generate mid-price and volume time-series for correlated stocks. We then show how our GAN architecture can be incorporated into state-of-the-art
diffusion models to enhance the quality of generated correlation structures. Specifically, the GAN's Critic serves as a quality evaluation module that guides the
diffusion process, enforcing learned correlation structures in the generated time-series. Our framework offers a lightweight and responsive solution for realistic stock
market simulation, explicitly modeling inter-asset correlation structures. We experimentally validate our framework against leading generative architectures, showing that it more effectively captures the stylized facts of stock
markets and models inter-asset correlations.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.CV
Equivalent Flows, Unequal Learning: Clean-Latent Prediction in Transformers
Funing Fu, Tenghui Wang, Guanyu Zhou et al. (+2 more)
Flow samplers consume velocity, but the neural network can predict the clean endpoint and convert it to velocity through a fixed affine readout. We study this choice with JLT, a latent
Transformer in a frozen variational autoencoder (VAE) representation. For squared error, the optimal clean and velocity predictors are algebraically equivalent; a finite
Transformer assigns different computation to its learned output under the two interfaces. A local Gaussian analysis identifies a known residual response supplied by the readout and isotropic target variance added by velocity prediction. Measured FLUX.2 channel spectra support this
geometric distinction: 90% of target variance occupies 83 of 128 clean directions versus 109 velocity directions. Under a matched velocity objective, clean prediction improves ImageNet FID-50K from 6.56 to 2.70 at Base scale and from 2.12 to 1.47 at Large scale, with lower FID at every measured Large checkpoint. Scaling clean prediction to 951M parameters reaches FID-50K 1.19 and IS 271.96. In addition, an objective ablation at Base scale shows that direct clean regression reaches FID-50K 2.38 without time-dependent error weighting. These results show how moving known computation outside the network changes learning under algebraically equivalent flow interfaces. Code: https://github.com/akatsuki-neo/JLT/blob/main/README.md
Read the full abstract ▾
All authors Funing Fu, Tenghui Wang, Guanyu Zhou, Junyong Cen, Qichao Zhu
Machine Learning
Mildly Overparameterized ReLU Networks on Orthogonal Data: Incremental Learning and Implicit Bias
James Town, Etienne Boursier, Ben Lewis et al. (+2 more)
The successful training of neural networks hinges on the use of first order optimization methods, yet the theoretical characterization of these methods remains incomplete. This is especially true in settings with mild overparameterization. In this work, we study the gradient flow dynamics of two-layer ReLU networks from small initialization with orthogonal training data. We prove the limiting flow converges to a saddle-to-saddle jump process as the initialization scale tends to zero, revealing an incremental learning phenomenon in which a new
neuron activates at each saddle. This analysis recovers the known result of Dana et al. (2025, arXiv:2502.16977) that the network interpolates the training data with high probability as soon as $m \gtrsim \log(n)$, where $m$ is the network width and $n$ is the number of training samples. This incremental process characterization also allows us to derive a novel implicit bias result: the learned interpolator has a squared $\ell_2$-norm scaling as $\sqrt{n}$, which is within a constant factor of the minimal $\ell_2$-norm interpolator. More broadly, our work provides the first rigorous proof of an incremental learning process for ReLU networks, whilst suggesting mildly overparameterized networks can converge to interpolating solutions whose complexity is of the same order as that of the optimal interpolator.
Read the full abstract ▾
All authors James Town, Etienne Boursier, Ben Lewis, Matthias Englert, Ranko Lazic
Machine Learning
Zhenglin Wan, Jingxuan Wu, Xingrui Yu et al. (+5 more)
Learning from demonstrations in
embodied control is often cast as behavioral cloning, and recent
diffusion or flow-matching policies improve this paradigm by modeling multi-modal expert actions. Yet these methods remain offline supervised learners: the policy is trained only on expert states and receives no corrective signal on the states it actually visits. On-policy
distillation (OPD) offers a natural remedy, but standard OPD assumes a strong fixed teacher, which is unavailable in demonstration-only control. We propose FA-OPD, an
adversarial dual on-policy
distillation method in which a Flow Matching (FM) teacher is learned from demonstrations and co-trained with a lightweight MLP student. The teacher provides two complementary signals on student rollouts. The reward channel learns an expert-likeness objective over state-action pairs and drives online exploration through long-horizon policy optimization. The action channel supplies dense local targets at student-visited states, stabilizing exploitation. FA-OPD couples them so that reward
distillation enables generalization beyond point-wise demonstrations, while action
distillation keeps exploration anchored near expert-like behavior. Across six
robot navigation, manipulation, and locomotion
benchmarks, FA-OPD beats strong baselines and shows much stronger robustness under noisy or limited demonstrations. Source code: https://github.com/vanzll/FA-OPD.
Read the full abstract ▾
All authors Zhenglin Wan, Jingxuan Wu, Xingrui Yu, Chubin Zhang, Mingcong Lei, Bo An, Ivor W. Tsang, Yang You
Machine Learning · Cross-listed from stat.ML
Gaussian Process-based learning with new MCMC-based implementation of Wishart prior on correlation matrix
Kane Warrior and Dalia Chakrabarty
In probabilstic supervised learning of an input-output relationship - as a sample function of a Gaussian Process (GP) - priors are typically specified for the hyperparameters of the kernel that parametrises the covariance function of the GP, where the induced covariance
matrix of the (resulting multivariate Normal) likelihood, governs the learning and prediction. When the sought function is highly multivariate, multiple lengthscale parameters must be learnt simultaneously, making inference difficult. We develop a “self-assembled” Wishart prior for the covariance
matrix, while undertaking
Bayesian inference on the kernel hyperparameters using MCMC. The construction uses a look-back window over recent MCMC iterations to define a time-step dependent scale
matrix, thereby introducing adaptiveness to the chain. Results suggest that direct prior specification on the covariance
matrix can be useful for diagnosing weakly informative inputs within the GP-based learning paradigm. We support our prior development with two distinct empirical illustrations - one on synthetic data, and another on a real-world
dataset.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.CL
LLMs Are Already Good Tutors: Training-Free Prompt Optimization for Pedagogical Math Tutoring
Unggi Lee, Minchul Shin, Yeil Jeong et al. (+5 more)
Aligning LLMs for math tutoring typically requires RL-based training with multi-GPU infrastructure. We investigate whether training-free
prompt optimization-evolving only the system
prompt via API calls-can serve as a practical alternative. We adapt 7 published methods and propose 5 education-specialized methods, evaluating these 12 methods under 5 conditions on 2 OOD
benchmark suites. All 12 best-per-method configurations surpass the strongest RL-trained baseline (R_total = 0.633), and our ParetoGrad achieves the best Pareto balance across post-test solve rate, leak control, and helpfulness, rather than dominating any single component. Behavioral analysis with an 82-code educational codebook reveals that training-free methods rely on teaching-knowledge patterns at 2-3x the rate of RL-trained models, with a compensating ~10 percentage-point reduction in intent-level scaffolding. We also find a task-dependent
reasoning mode effect consistent across training-free and RL-based paradigms. Our approach enables efficient development of pedagogically aligned LLM tutors with
prompts alone and minimal compute.
Read the full abstract ▾
All authors Unggi Lee, Minchul Shin, Yeil Jeong, Sookbun Lee, Jeongsu Moon, Kyungtae Joo, Eunjoo Lee, Hoilym Kwon
Machine Learning
ReMoE: Boosting Expert Reuse through Router Fine-Tuning in Memory-Constrained MoE LLM Inference
Xiongwei Zhu, Xiaojian Liao, Tianyang Jiang et al. (+3 more)
Fine-grained Mixture-of-Experts (MoE) models sparsely activate only a subset of experts per
token, reducing activated computation while maintaining high model capacity. However, in
memory-constrained inference scenarios, only a small set of experts can be cached. Experts not in the cache must be fetched from slow external storage (e.g., UFS), leading to frequent evictions and substantial I/O overhead. We propose ReMoE, a router fine-tuning framework designed to boost
token-wise expert reuse. ReMoE biases the router toward recently selected experts, producing temporally stable routing that better matches cache locality constraints. By increasing short-horizon expert reuse, ReMoE reduces expert fetches from storage without adding inference-time computation. Experiments on DeepSeek and Qwen models show that ReMoE improves expert reuse by 26% while maintaining downstream task performance. Real-system evaluations further confirm these benefits, improving output throughput by 8.4% under vLLM GPU-CPU expert offloading and reducing TPOT by 43.6-49.8% under llama.cpp on Jetson Orin NX, corresponding to a 1.77-1.99$\times$ decode speedup across diverse workloads. Checkpoints and usage instructions are available at https://github.com/BUAA-OSCAR/ReMoE.
Read the full abstract ▾
All authors Xiongwei Zhu, Xiaojian Liao, Tianyang Jiang, Yusen Zhang, Liang Wang, Limin Xiao
Machine Learning
Trust Region Q Adjoint Matching
Yonghoon Dong, Kyungmin Lee, Changyeon Kim et al. (+2 more)
Off-policy reinforcement learning of pretrained flow policies remains challenging due to the instability of optimization arising from the multi-step sampling process. Recently, Q-learning with Adjoint Matching (QAM) addressed this by recasting policy improvement as a
stochastic optimal control (SOC) problem guided by a learned critic. However, QAM inherits a fundamental fragility of critic-guided improvement, since small critic errors can be exponentially amplified and often lead to performance collapse. This paper introduces Trust Region Q Adjoint Matching (TRQAM), a stable off-policy fine-tuning
algorithm that adaptively controls the path-space KL between the fine-tuned and pretrained policies through projected dual descent. Specifically, we adapt a trust-region parameter $\lambda$ inside the sampling process of the flow policy and prove that $\lambda$ exactly weights the path-space KL in the SOC objective. As a result, our method can tightly control the deviation from the pretrained policy within a prescribed KL budget, achieving stable off-policy RL. Through experiments on 50 OGBench tasks, TRQAM consistently outperforms prior methods in both offline RL and offline-to-online RL. In particular, TRQAM achieves an overall success rate of 68% in offline RL, substantially improving over the strongest baseline at 46%.
Read the full abstract ▾
All authors Yonghoon Dong, Kyungmin Lee, Changyeon Kim, Jaehyuk Kim, Jinwoo Shin
Machine Learning
Two Speeds of Learning: A Representation-Readout Decomposition of Grokking and Double Descent
Chi-Ning Chou, Oscar Uzdelewicz, Neng-Chun Chiu et al. (+2 more)
Training loss and accuracy are the standard signals used to monitor generalization during deep neural network training. Two well-documented phenomena complicate this picture: in grokking, train loss falls rapidly while test performance improves abruptly only after a long delay; in epoch-wise double descent, train loss decreases monotonically while test loss or error rises and falls. Existing accounts are often task-specific, and a task-agnostic analysis framework for diagnosing and explaining these phenomena across realistic tasks and architectures is missing. We address this challenge by analyzing two competing processes that underlie learning dynamics: representation learning in the encoder and readout calibration in the final classifier. Using tools from representational geometry, neural tangent kernels, and linear probing, we show that both processes are active throughout training, with the fluctuations of their relative speed giving rise to seemingly anomalous generalization dynamics. Applying the representation-readout decomposition to grokking across a wide range of tasks and architectures, we find that the readout is train-biased before grokking onset, and representation learning is gradual but not absent, contrary to the lazy-to-rich account. The framework further provides diagnostic signatures distinguishing spurious from genuine generalization: in a previously reported MNIST grokking example and an epoch-wise double descent example, apparent delayed or non-monotone generalization is shown to arise from representation degradation and readout misalignment induced by non-standard training recipes. Together, these results establish the representation-readout decomposition as a top-down framework for understanding learning dynamics and revealing underlying
algorithms for interpretability research.
Read the full abstract ▾
All authors Chi-Ning Chou, Oscar Uzdelewicz, Neng-Chun Chiu, Yao-Yuan Yang, SueYeon Chung
Machine Learning · Cross-listed from cs.MA
Cost of Structural Learning Under Censored Feedback: A Threshold-Bandit Approach
Michael Ledford and William Regli
In many multi-
agent applications, tasks yield rewards only when executed by a coalition meeting an unknown size threshold; otherwise, feedback is fully censored. This censorship creates an identifiability problem:
agents cannot distinguish
stochastic failure from insufficient coordination. We formalize this setting as the Threshold-Activated Cooperative Multi-Armed Bandit (TAC-MAB) and analyze it under both centralized and decentralized coordination. We show that a centralized
algorithm (C-TAC) achieves cumulative regret O(log T), decomposed into a structural-search term that captures the cost of resolving feasibility under censored feedback and a statistical-monitoring term for value estimation. We then introduce D-TAC, a decentralized event-triggered protocol in which
agents synchronize only when their structural beliefs change. Empirically, D-TAC achieves a 23x reduction in communication relative to the centralized baseline while preserving feasibility
alignment under conservative belief fusion. These results characterize the coordination cost of learning under censored feedback and show that near-centralized communication efficiency is achievable without continuous synchronization.
Read the full abstract ▾
Machine Learning
Learning to Orchestrate Agents under Uncertainty
Mary Chriselda Antony Oliver, Lan Jiang, Aaron Bundi Anampiu et al. (+3 more)
Adaptive orchestration of heterogeneous
agents requires making sequential delegation decisions under uncertain and evolving
agent behaviour, e.g., coordinating specialised AI models with varying reliability, cost, and response quality. While prior work on
agent orchestration focuses on performance or cost, uncertainty in
agent reliability and output distributions is typically not modelled explicitly at the orchestration level. In this work, we study the problem of adaptive orchestration of heterogeneous
agents under uncertainty, where a meta-controller must decide when to delegate to an
agent, accounting for reliability, cost, and uncertainty. We propose BOT-Orch, a lightweight framework that recasts orchestration as a bandit problem over
agents, regularized by OT distances between
agent output distributions and task-specific reference distributions. We show that the regularised orchestration enjoys $\mathcal{O}(\sqrt{T})$ regret under standard assumptions, and provably induces preference ordering among
agents with identical mean rewards but differing distributional
alignment. Empirically, we demonstrate that BOT-Orch outperforms standard bandit and heuristic baselines in synthetic but
adversarial task allocation settings with heterogeneous, non-i.i.d.
agent behaviour.
Read the full abstract ▾
All authors Mary Chriselda Antony Oliver, Lan Jiang, Aaron Bundi Anampiu, Elaf Almahmoud, Francesco Quinzan, Umang Bhatt
Machine Learning
Learning Dynamic Graph Representations through Timespan View Contrasts
Yiming Xu, Zhen Peng, Bin Shi et al. (+2 more)
The rich information underlying graphs has inspired further investigation of unsupervised graph representation. Existing studies mainly depend on node features and topological properties within static graphs to create self-supervised signals, neglecting the temporal components carried by real-world graph data, such as timestamps of edges. To overcome this limitation, this paper explores how to model temporal
evolution on dynamic graphs elegantly. Specifically, we introduce a new inductive bias, namely temporal translation invariance, which illustrates the tendency of the identical node to keep similar labels across different timespans. Based on this assumption, we develop a dynamic graph representation framework CLDG that encourages the node to maintain locally consistent temporal translation invariance through contrastive learning on different timespans. Except for standard CLDG which only considers explicit topological links, our further proposed CLDG++ additionally employs graph
diffusion to uncover global contextual correlations between nodes, and designs a multi-scale contrastive learning objective composed of local-local, local-global, and global-global contrasts to enhance representation capabilities. Interestingly, by measuring the consistency between different timespans to shape anomaly indicators, CLDG and CLDG++ are seamlessly integrated with the task of spotting anomalies on dynamic graphs, which has broad applications in many high-impact domains, such as finance, cybersecurity, and healthcare. Experiments demonstrate that CLDG and CLDG++ both exhibit desirable performance in downstream tasks including node classification and dynamic graph anomaly detection. Moreover, CLDG significantly reduces time and space complexity by implicitly exploiting temporal cues instead of complicated sequence models.
Read the full abstract ▾
All authors Yiming Xu, Zhen Peng, Bin Shi, Xu Hua, Bo Dong
Machine Learning · Cross-listed from cs.CL
FalAR: A Large-scale Speaker-Annotated European Portuguese Speech Corpus of Parliamentary Sessions
Francisco Teixeira, Carlos Carvalho, Mariana Julião et al. (+7 more)
State-of-the-art performance for Automatic
Speech Recognition (ASR) largely depends on the availability of large-scale labeled corpora. This creates a demand for increased data collection efforts, particularly for under-represented
languages and dialectal varieties. Due to having considerably fewer speakers (around 11 million), European Portuguese (EP) is overshadowed by Brazilian Portuguese (BP) (around 200 million speakers) in currently available large-scale
speech data resources, resulting in under-performing
speech-based systems for EP users. To address this gap, and following similar data collection efforts for other
languages, we present FalAR, a large-scale, speaker-annotated
speech corpus of European Portuguese parliamentary sessions. Spanning approximately 20 years, FalAR comprises 5,800 hours of
speech data. In addition, 4,850 hours have speaker identity annotations, for a total of 1,180 speakers with associated metadata including age, gender, political affiliation, and parliamentary role. The corpus was built using a state-of-the-art EP CAMÕES ASR model for transcription-reference
alignment. In this paper, we describe the data collection process, together with the main characteristics of the FalAR corpus. Furthermore, we evaluate the trade-off between data quantity and
alignment accuracy on ASR performance, with our experiments demonstrating that incorporating FalAR as pre-training data yields up to 14% relative WER improvement over baseline models.
Read the full abstract ▾
All authors Francisco Teixeira, Carlos Carvalho, Mariana Julião, Catarina Botelho, Rubén Solera-Ureña, Sérgio Paulo, Thomas Rolland, Ben Peters, Isabel Trancoso, Alberto Abad
Machine Learning · Cross-listed from cs.CL
BhashaSetu: A Data-Centric Approach to Low-Resource Machine Translation
Param Thakkar, Anushka Yadav, Michael Tiemann et al. (+3 more)
We present BhashaSetu, a linguistically enriched English–Marathi parallel
dataset addressing persistent data limitations in low-resource neural machine translation (NMT). Marathi, spoken by over 95 million people, remains underrepresented in high-quality parallel corpora across diverse domains. Our
dataset comprises 2.78 million sentence pairs from heterogeneous sources including news, politics, healthcare, literature, and culture, with stemmed and lemmatized representations to support morphology-aware analysis. We
benchmark multiple state-of-the-art translation models using BLEU, spBLEU, chrF++, and TER metrics, and conduct parameter-efficient fine-tuning of NLLB-200-distilled-600M using LoRA. A key finding from our ablation: corpus-level deduplication is the single largest preprocessing contributor to downstream quality (removing it reduces performance by 1.17 BLEU and 2.21 chrF++), demonstrating that disciplined cross-source corpus hygiene is a low-cost, high-impact intervention for low-resource, morphologically rich
languages. The
dataset is publicly released to promote reproducible and linguistically informed low-resource NMT research.
Read the full abstract ▾
All authors Param Thakkar, Anushka Yadav, Michael Tiemann, Abhi Mehta, Akshita Bhasin, Shrinivas Khedkar
Machine Learning · Cross-listed from stat.ML
Causal Representation Learning for Generalisable Recommendation
Yorgos Felekis, Michael O'Riordan, Oriol Corcoll et al. (+1 more)
Predictive models trained on observational data often fail to generalise when deployed, especially when the training data is a product of the system being optimised. Recommender systems are a canonical example: they are trained on interaction logs confounded by the deployed policy, past user behaviour, and platform filtering. As a result, the training distribution differs from the candidate distribution scored at serving time, a gap that makes offline metrics unreliable predictors of online performance. We address the distribution shift problem with a method motivated by causal representation learning (CRL). We propose an information-theoretic disentanglement criterion and prove that its optimum excludes confounder-dependent non-causal information from the treatment representation. We derive a tractable variational lower bound making the criterion optimisable from finite observational data. The scope of our method is narrower than that of much of the CRL literature, in that we target better generalisation under distribution shift, not full identification of latent causal factors. This narrower target makes the method practical, requiring only existing confounded logs, applying to any standard supervised model, and adding no inference-time cost. Our headline evaluation is an A/B test with millions of users on Spotify, applied to a production ranker for personalised playlist generation. A CRL variant matched in capacity to the production baseline performed on par offline but delivered substantial online gains in listener engagement. Complementary evidence on a synthetic
dataset with known causal structure and a semi-synthetic variant of the public KuaiRand recommendation
dataset shows similar in-distribution performance to the baseline but gains under distribution shift. Across all three settings, adding our causal disentanglement objective yields better out-of-distribution generalisation.
Read the full abstract ▾
All authors Yorgos Felekis, Michael O'Riordan, Oriol Corcoll, Ciarán M. Gilligan-Lee
Machine Learning · Cross-listed from cs.CL
Tracing Computation Density in LLMs
Corentin Kervadec, Iuliia Lysova, Iuri Macocco et al. (+2 more)
Transformer-based large
language models (LLMs) are comprised of billions of parameters arranged in deep and wide computational graphs, but it is not clear that they exploit their full capacity for all inputs. We introduce the s-Trace method to efficiently estimate a subgraph of size s that approximates a full model output. With this method, we find the computation in a variety of LLMs to be organized in two distinct phases. A small subgraph mostly composed of early-layer nodes can reconstruct the head of the full model output distribution. Adding further nodes, mostly located in later layers and increasingly consisting of
attention heads, leads to incremental refinements in approximating the full output distribution. We find moreover that the amount of necessary computation per input correlates with model uncertainty, and that sparser subgraphs encode shallow statistics, such as unigram frequency. Overall, our results suggest a consistent modular organization in effective LLM computation, with a sparse early-layer core providing a rough prediction that is further refined through denser computations in later layers.
Read the full abstract ▾
All authors Corentin Kervadec, Iuliia Lysova, Iuri Macocco, Marco Baroni, Gemma Boleda
Machine Learning
Less is More: Early Stopping Rollout for On-Policy Distillation
Zhou Ziheng, Jiaqi Li, Huacong Tang et al. (+2 more)
On-policy
distillation has recently emerged as a promising alternative to standard sequence-level imitation, training a student by scoring its own rollouts with a teacher model. However, we observe “Off-policy Teacher Decay” problem in this paradigm: for the later
tokens, with student's earlier trajectory as context that is off-policy to the teacher, the teacher's ability to produce a corrective score would decay, and may fall back to
token-completion behavior learned in the pre-training stage. We empirically verify this problem, and we propose Early Stopping Rollout (ESR) to fix it: a simple yet effective
distillation strategy that simply restricts the rollout generation to the first response
tokens. We show that ESR both surpasses the full rollout OPD performance across model size, family, tasks and training regime, and exhibit much higher GPU efficiency and training stability, especially under cross model family scenarios. We further investigate the mechanism behind this surprising performance and discovered "Cascading
Alignment" and "Sub-mode Commitment" effect of ESR that may explain why it works effectively and even sometimes exceeding the teacher model performance. Besides, we show that this position-based
token selection strategy cannot be fully explainable by KL divergence and
entropy signals.
Read the full abstract ▾
All authors Zhou Ziheng, Jiaqi Li, Huacong Tang, Ying Nian Wu, Demetri Terzopoulos
Machine Learning
SQARL: A Size-Agnostic Reinforcement Learning approach for Circuit Allocation in Distributed Quantum Architectures
Víctor Carballo, Júlia López-Closa and Mario Martin
The scaling of
quantum processors is currently limited by technical challenges such as decoherence and cross-talk. As the number of
qubits grows, interference increases the computational noise. Distributed
quantum computing addresses these limitations by interconnecting smaller, easier-to-handle
quantum processors (cores), but it introduces the challenge of minimizing slow, error-prone inter-core communication. The task of distributing
quantum circuits across cores while minimizing communication costs is known as the
Qubit Allocation problem. This work focuses on developing a deep learning approach to this problem, emphasizing flexibility to
quantum hardware
topology and improving state-of-the-art performance. Heuristic and non-learning
algorithms, such as the Hungarian
Qubit Allocation (HQA), currently represent the state of the art. Reinforcement Learning (RL) approaches leverage learned allocation policies but often lack flexibility, requiring retraining when hardware configurations change, and they fall short of the solution quality achieved by non-learning methods. However, learning mechanisms could outperform human-crafted heuristics. To overcome these limitations, this work proposes a flexible,
transformer-based architecture that can handle arbitrary numbers of
qubits and cores without retraining. Results show that the trained policy consistently outperforms the previous RL state of the art and narrows the gap between RL and HQA for the most common
circuits. It achieves a 33% reduction in allocation cost relative to the HQA for the Cuccaro Adder and 25% on average for random
circuits. These findings show that learning-based approaches can effectively match the performance of hand-crafted heuristics, a crucial step towards their application in real-world scenarios.
Read the full abstract ▾
Machine Learning · Cross-listed from cs.CL
How Do Uncertainty Estimators Align with Hallucination Occurrence in LLMs?
Yedidia Agnimo, Anna Korba, Annabelle Blangero et al. (+2 more)
Large
language models (LLMs) can hallucinate, but do model uncertainty
estimators reliably reveal when they do? Although uncertainty
estimators are increasingly used to identify unreliable generations, their relationship with
hallucination remains poorly understood and is often taken for granted. We systematically evaluate 46 uncertainty
estimators across complementary
hallucination settings, covering both intrinsic
hallucinations, where outputs are unsupported by the provided context, and extrinsic
hallucinations, where claims are unsupported by expected model knowledge. Across four
benchmarks and multiple LLMs, we find that the association between uncertainty and
hallucination varies widely and is often weak, depending on the
hallucination setting and the model under evaluation. These results challenge the use of uncertainty as a direct signal of
hallucination and clarify when it provides useful information.
Read the full abstract ▾
All authors Yedidia Agnimo, Anna Korba, Annabelle Blangero, Nicolas Chesneau, Karteek Alahari
Machine Learning
SCENT: Aligning Mass Spectra with Molecular Structure for Olfactory Perception
Ziqi Zhang, Eunyeong Jin, Miguel Vasco et al. (+6 more)
Predicting human olfactory perception from molecular structure has seen remarkable progress, yet these approaches require explicit chemical structure at inference, which is not available in practical sensing settings. We address this gap by exploring direct electron ionization mass spectrometry (EI-MS), a sensing technique that acquires chemically informative fragmentation fingerprints in seconds, as an alternative input modality for olfactory prediction. We contribute Spectrum-to-Chemical
Embedding alignmeNT (SCENT), a multi-modal contrastive learning framework that aligns EI-MS representations with pretrained chemical structure
embeddings, while requiring only mass spectra at inference. On the multi-label odor descriptor prediction task, SCENT significantly outperforms MS-only baselines and achieves performance comparable to structure-based models, despite requiring no explicit molecular structure at test time. The learned representations also better approximate continuous human perceptual ratings and generalize to real-world lab-measured spectra, suggesting that cross-modal
alignment is an effective strategy for grounding analytical spectra in chemical semantics.
Read the full abstract ▾
All authors Ziqi Zhang, Eunyeong Jin, Miguel Vasco, Farzaneh Taleb, Nona Rajabi, Alexandra Gutmann, Jonathan Williams, Antônio H. Ribeiro, Danica Kragic
Machine Learning
Sampling Data with Chains of Forward-Backward Diffusion Steps
Hyunmo Kang, Noam Itzhak Levi, Corinna Elena Wegner et al. (+2 more)
Sampling from learned high-dimensional distributions is a foundational computational problem. We introduce U-turn chains: Markov chains obtained by iterating short forward-backward steps of a
diffusion model, in which each step proposes a move that remains on the learned data
manifold and, paired with a Metropolis-Hastings correction, samples from energy-modified targets. For synthetic
languages, we show that minimal U-turn dynamics undergoes an ergodicity-breaking phase transition driven by fragmentation of the data
manifold; ergodicity is restored at larger U-turn magnitude. In the non-ergodic regime, low-level features relax faster than high-level ones, an ordering that inverts only at sufficiently large U-turn magnitude. We test these predictions on natural
language and natural images. In both modalities, minimal U-turns relax slowly, especially for high-level features approximated by deep representations in CNNs or LLMs. The layer-ordering inversion appears only at large noise when mixing is efficient – signatures consistent with strongly constrained, weakly mixing local dynamics. We discuss the implications of these results for sampling with
diffusion models.
Read the full abstract ▾
All authors Hyunmo Kang, Noam Itzhak Levi, Corinna Elena Wegner, Daniel J. Korchinski, Matthieu Wyart
Machine Learning
Probabilistic Recurrent Intention Switching Model
Wenyuan Sheng, Hao Zhu and Joschka Boedecker
Inverse reinforcement learning (IRL) recovers reward functions from observed behavior, yet traditional methods assume a single stationary reward that cannot capture goal switching within an episode. Recent multi-intention IRL methods address this by segmenting trajectories, but model intention transitions as either a memoryless Markov chain or via manual state augmentation with a fixed history window. We propose the Probabilistic Recurrent Intention Switching Model (PRISM), which replaces both mechanisms with a lightweight recurrent network that maps observation history to a per-step intention distribution. We prove that the resulting EM objective decomposes exactly into independent per-intention reward subproblems, each solvable in closed form, yielding an $\mathcal{O}(nK)$ E-step with no variational approximation. We evaluate PRISM on a non-Markovian gridworld, a mouse labyrinth, and BridgeData~V2 robotic manipulation, the first large-scale robotic application of multi-intention IRL. Across all settings PRISM achieves the highest held-out log-likelihood while recovering nameable, temporally coherent intentions from unlabeled demonstrations, suggesting that discrete goal switching is present in both biological and artificial
agents.
Read the full abstract ▾