Recent work has shown that large language models (LLMs) can enhance recommender systems by integrating collaborative filtering (CF) signals through hybrid prompting. However, most existing CF-LLM frameworks collapse explicit ratings into implicit or positive-only feedback, discarding the ordinal structure that conveys fine-grained preference strength. As a result, these models struggle to exploit graded semantics and nuanced preference distinctions. We propose Ordinal Semantic Anchoring (OSA), a hybrid CF-LLM framework that explicitly incorporates preference strength by modeling interaction-level user feedback. OSA represents ordinal preference levels as numeric textual tokens and uses their token embeddings as semantic anchors to align user-item interaction representations in the LLM latent space. Through strength-aware alignment across ordinal levels, OSA preserves preference semantics when integrating collaborative signals with LLMs. Experiments on multiple real-world datasets demonstrate that OSA consistently outperforms existing baselines, particularly in pairwise preference evaluation, highlighting its effectiveness in modeling fine-grained user preferences over prior CF-LLM methods.

With the increasing demand for histopathological specimen examination and diagnostic reporting, Multiple Instance Learning (MIL) has received heightened research focus as a viable solution for AI-centric diagnostic aid. Recently, to improve its performance and make it work more like a pathologist, several MIL approaches based on the use of multiple-resolution images have been proposed, delivering often higher performance than those that use single-resolution images. Despite impressive recent developments of multiple-resolution MIL, previous approaches only focus on improving performance, thereby lacking research on well-calibrated MIL that clinical experts can rely on for trustworthy diagnostic results. In this study, we propose Uncertainty-Focused Calibrated MIL (UFC-MIL), which more closely mimics the pathologists’ examination behaviors while providing calibrated diagnostic predictions, using multiple images with different resolutions. UFC-MIL includes a novel patch-wise loss that learns the latent patterns of instances and expresses their uncertainty for classification. Also, the attention-based architecture with a neighbor patch aggregation module collects features for the classifier. In addition, aggregated predictions are calibrated through patch-level uncertainty without requiring multiple iterative inferences, which is a key practical advantage. Against challenging public datasets, UFC-MIL shows superior performance in model calibration while achieving classification accuracy comparable to that of state-of-the-art methods.

We consider the universal discrete filtering problem, where an input sequence generated by an unknown source passes through a discrete memoryless channel, and the goal is to estimate its components based on the output sequence, with limited lookahead or delay. We propose and establish the universality of a family of schemes for this setting. These schemes are induced by universal Sequential Probability Assignments (SPAs), and inherit their computational properties. We show that the schemes induced by LZ78 (a Lempel-Ziv compression algorithm) are practically implementable and well-suited for scenarios with limited computational resources and latency constraints. As a byproduct of our analysis, we obtain novel upper and lower bounds in the purely Bayesian setting using some of the intermediate results.

We consider the problem of non-coherent over-the-air computation (AirComp), where$n$devices carry highdimensional data vectors$\mathrm{x}_{i} \in \mathbb{R}^{d}$of sparsity$\left\vert\mathrm{x}_{i}\right\vert_{0} \leq k$and the sum of these data vectors has to be computed at a receiver. Previous results on non-coherent AirComp require more than$d$channel uses to compute functions of$\mathrm{x}_{i}$, where the extra redundancy is used to combat non-coherent signal aggregation. However, if the data vectors are sparse, sparsity can be exploited to offer significantly cheaper communication. In this paper, we propose to use random transforms to transmit lower-dimensional projections$s_{i} \in \mathbb{R}^{T}$of the data vectors. These projected vectors are communicated to the receiver using a majority vote (MV)AirComp scheme, which estimates the bit-vector corresponding to the signs of the aggregated projections, i.e.,$\mathbf{y}=\text{sign}\left(\sum_{i} \mathbf{s}_{i}\right)$. By leveraging 1-bit compressed sensing (1bCS) at the receiver, the real-valued and high-dimensional aggregate$\sum_{i} \mathrm{x}_{i}$can be recovered from$y$. We prove analytically that the proposed MVCS scheme estimates the aggregate data vector$\sum_{i} \mathrm{x}_{i}$with$\ell_{2}$-norm error$\epsilon$in$T=\mathcal{O}\left(k n \log (d) / \epsilon^{2}\right)$channel uses. We consider distributed histogram estimation, a canonical building block for federated analytics, as an aplication for MVCS where the data vectors$\mathrm{x}_{i}$are inherently 1 -sparse. Our numerical evaluations demonstrate that our scheme achieves the same order of communication cost as state-of-the-art methods while avoiding the complexity and overhead of additional cryptographic tools.

Millimeter wave (mmWave) communications have been recognized as an important technology to address the growing data demand of next-generation wireless communications. Since mmWave communications have the inherent disadvantage of severe signal attenuation, the beamforming technique using a large antenna array is widely adopted at the base station (BS). For achieving optimal beamforming performance, a beam management process for angular channel information must be carried out. However, conventional beam management techniques suffer from long latency and inaccuracies in determining the beam direction due to the quantized codebook design. To overcome this issue, we propose an active receiver-based beam management (AR-BM) in which the user equipment (UE) determines and reports the grid-free channel direction without the need for additional refinement stages, while also achieving more accuracy. In our scheme, BS transmits wide training beams generated by partially activated arrays. Then, UE determines and reports gridfree direction by updating the estimated direction comparing the measured reference signal received power (RSRP) and the function-derived RSRP. Numerical results show that the proposed scheme achieves a$\mathbf{1 4 \%}$improvement in the data rate and a$\mathbf{9 0 \%}$reduction in the latency.

We consider the universal discrete filtering problem, where an input sequence generated by an unknown source passes through a discrete memoryless channel, and the goal is to estimate its components based on the output sequence, with limited lookahead or delay. We propose and establish the uni-versality of a family of schemes for this setting. These schemes are induced by universal Sequential Probability Assignments (SPAs), and inherit their computational properties. We show that the schemes induced by LZ78 are practically implementable and well-suited for scenarios with limited computational re-sources and latency constraints.

We introduce a lightweight and accurate localization method that only utilizes the geometry of 2D-3D lines. Given a pre-captured 3D map, our approach localizes a panorama image, taking advantage of the holistic 360° view. The system mitigates potential privacy breaches or domain discrepancies by avoiding trained or hand-crafted visual descriptors. However, as lines alone can be ambiguous, we express distinctive yet compact spatial contexts from relationships between lines, namely the dominant directions of parallel lines and the intersection between non-parallel lines. The resulting representations are efficient in processing time and memory compared to conventional visual descriptor-based methods. Given the groups of dominant line directions and their intersections, we accelerate the search process to test thousands of pose candidates in less than a millisecond without sacrificing accuracy. We empirically show that the proposed 2D-3D matching can localize panoramas for challenging scenes with similar structures, dramatic domain shifts or illumination changes. Our fully geometric approach does not involve extensive parameter tuning or neural network training, making it a practical algorithm that can be readily deployed in the real world. Project page including the code is available through this link: https://82magnolia.github.io/fgpl/.

We consider the problem of secure histogram es-timation, where$n$users hold private items xifrom a size-d domain and a server aims to estimate the histogram of the user items. Previous results utilizing orthogonal communication schemes have shown that this problem can be solved securely with a total communication cost of O(n2log(d)) bits by hiding each item xiwith a mask. In this paper, we offer a different approach to achieving secure aggregation. Instead of masking the data, our scheme protects individuals by aggregating their messages via a multiple-access channel. A naive communication scheme over the multiple-access channel requires$d$channel uses, which is generally worse than the O(n21og(d)) bits communication cost of the prior art in the most relevant regime$d$>>$n$. Instead, we propose a new scheme that we call Over-the-Air Group Testing (AirG T) which uses group testing codes to solve the histogram estimation problem in O(n log(d)) channel uses. AirGT reconstructs the histogram exactly with a vanishing probability of error Perror= O(d-T) that drops exponentially in the number of channel uses$T$.

Biped controller design pursues two fundamental goals; simulated walking should look human-like and robust against perturbation while maintaining its balance. Normal gait is a pattern of walking that humans normally adopt in undisturbed situations. It has previously been postulated that normal gait is more energy efficient than abnormal or impaired gaits. However, it is not clear whether normal gait is also superior to abnormal gait patterns with respect to other factors, such as stability. Understanding the correlation between gait and stability is an important aspect of biped controller design. We studied this issue in two sets of experiments with human participants and a simulated biped. The experiments evaluated the degree of resilience to external pushes for various gait patterns. We identified four gait factors that affect the balance-recovery capabilities of both human and simulated walking. We found that crouch gait is significantly more stable than normal gait against lateral push. Walking speed and the timing/magnitude of disturbance also affect gait stability. Our work would provide a potential way to compare the performance of biped controllers by normalizing their output gaits and improve their performance by adjusting these decisive factors.