Papers

Fair clustering aims to prevent algorithmic bias against protected groups. We study Euclidean fair $k$-means, in which the proportion of each protected group in every cluster lies between prescribed lower and upper bounds. Because these constraints violate the nearest-center assignment property, center selection is coupled with point assignment. We introduce the drift-regularized relaxation, a new convex relaxation solved by a modified Frank-Wolfe method with linear-programming iterations, whose optimality condition controls the cost of converting its solution into $k$ fair clusters. Combined with any $\rho$-approximation algorithm for weighted $k$-means, the relaxation yields an exactly fair fractional solution of cost at most $2\rho+1+\epsilon$ times the fractional optimum, improving the previous ratio $1+4\rho+\epsilon$. In fixed dimension, this improves the best known ratio from $5+O(\epsilon)$ to $3+O(\epsilon)$. The fractional solution can be rounded to an integral one without changing the centers or increasing the cost, at a bounded additive violation of the fairness bounds. We also show that a $(3-\delta)$-approximation is NP-hard for every fixed $\delta>0$, even on the line with fractional assignments, so $3$ is the exact approximation threshold for fractional fair $k$-means in fixed dimension. The same guarantees hold for $k$-sparse Wasserstein barycenters, a special case of fractional fair $k$-means. More generally, they hold whenever every cluster must satisfy given homogeneous linear inequalities in the fractions of its points, such as a bound on the ratio between two groups or on the average of a numerical attribute in every cluster.

When you merge two fine-tuned models from the same base checkpoint by simply averaging their weights, you implicitly assume their attention subspaces are still aligned. Recent work attempts to fix misalignments by learning an invertible correction matrix, $M$, for each model's query and key projections. This correction cancels out—using $M$ on the query side and $M^{-T}$ on the key side—right before the dot product. However, this cancellation is only exact if nothing sits between the projection and the dot product. In reality, almost all modern open-weight language models put a rotary position embedding (RoPE) exactly there. In this paper, we show that this cancellation is exact under RoPE if and only if $M$ commutes with RoPE's per-position rotation. We derive the specific class of matrices where this holds: a scaled rotation acting independently within each RoPE frequency pair. This forms a strict, low-dimensional subset of the unconstrained matrices that current methods normally train. Building on this, we turn this constrained matrix class into a new merging method. While keeping the base weights entirely frozen, two fine-tunes each learn their own RoPE-compliant correction matrices. We optimize these corrections against a chosen blend ratio so the final result can be adjusted post-hoc like a dial, rather than locked into a single fixed merge. Our default approach trains at one fixed blend ratio, similar to how LoRA sets its scaling hyperparameter in advance. We also experiment with resampling the blend ratio randomly at every training step, and we report the results of both approaches.

We describe geometric-function-atlas, our open-source Python package for the sharp extremal problems of geometric function theory. We organise it around a catalogue of thirty-nine Ma–Minda starlike generators. From this catalogue we compute exact Taylor coefficients, closed-form Fekete–Szegő constants, exact coefficients of the Ma–Minda extremal function, and admissibility screens. Our verifier answers membership questions for normalised polynomials at three levels of evidence: a floating-point grid screen, an exact sufficient condition decided in rational arithmetic, and a certified interval enclosure at the worst screened point. Every answer names the level at which we obtained it. We ship a checksummed artifact snapshot with three hundred and six coefficient certificates and seven hundred and two directed inclusion radii. Eight reviewed radius lanes carry certificates whose proof chains we replay symbolically, and we re-execute every coefficient certificate through our exact Schur-parameter machinery on request. We emit all results through one versioned envelope that records the method, the evidence status, the assumptions, and the artifact identifiers. Two optional laboratories apply the same discipline to cryptographic S-box metrics and to image-quality metrics. We present our design, state as propositions what each tier establishes, follow one radius lane from screen to replayed certificate, report measured timings, and place our package among symbolic-algebra, rigorous-numerics, and mathematical-database software. We release geometric-function-atlas under the MIT licence on the Python Package Index and at https://github.com/Prasanna28Devadiga/geometric-function-atlas.

Child speech differs from adult speech in acoustics, prosody, and linguistic structures. Speech disfluencies (such as repetitions) further challenge automatic understanding. While Audio Language Models (ALMs) show strong semantic reasoning from speech audio, their ability to reason about disfluent child speech in mixed-speaker settings remains unexplored. We investigate this through two tasks: child-focused semantic summarization and speech entailment. Experiments use recordings of children who stutter in mixed speaker interviews without explicit speaker separation. Models are instruction-guided to focus on the child, preserve clinically relevant disfluencies, and avoid adult-speech leakage. Evaluation combines LLM-based judges and reference-based metrics, anchored by transcript-oracle baselines to isolate errors. Results show that while ALMs extract high-level meaning from stuttered speech, reasoning degrades significantly with increased

Key–value (KV) cache compression is an effective way to reduce the memory overhead of large language model (LLM) inference, particularly for long-context workloads. However, existing compression methods make different trade-offs among accuracy, inference latency, and peak KV cache memory utilization, making a single fixed configuration unsuitable across different prompts and resource constraints. We introduce MetaKV, an adaptive framework that selects a KV cache compression configuration for each input prompt based on user-specified latency and peak memory budgets. MetaKV uses lightweight prediction models to estimate the end-to-end latency, peak memory, and probability of a correct response for each candidate configuration, and selects the configuration that best satisfies the latency-memory constraints while preserving accuracy. We evaluate MetaKV across ten configurations from three representative KV cache compression methods, KVQuant, H$_2$O, and RocketKV, together with an uncompressed FP16 configuration, on four datasets covering mathematics, science, commonsense reasoning, and reading comprehension. Across a wide range of latency and peak memory constraints, MetaKV consistently outperforms the best static configuration, improving constrained success rate (CSR), the fraction of prompts answered correctly while satisfying both constraints, by approximately 0.07 on average and up to 0.135. These results demonstrate the benefit of adapting KV cache compression to individual prompts and latency-memory constraints. Code is available at https://github.com/MichaelWang0505/MetaKV.git

Sign Language Translation has advanced with deep learning, yet evaluations remain largely signer-dependent, with overlapping signers across train/dev/test. This raises concerns about whether models truly generalise or instead rely on signer-specific regularities. We conduct signer-fold cross-validation on GFSLT-VLP, GASLT, and SignCL, three leading, publicly available, gloss-free SLT models, on CSL-Daily and PHOENIX14T. Under signer-independent evaluation, performance drops sharply: on PHOENIX14T, GFSLT-VLP falls from BLEU-4 21.44 to 3.59 and ROUGE-L 42.49 to 11.89; GASLT from 15.74 to 8.26; and SignCL from 22.74 to 3.66. We also observe that in CSL-Daily many target sentences are performed by multiple signers, so common splits can place identical sentences in both training and test, inflating absolute scores by rewarding recall of recurring sentences rather than genuine generalisation. These findings indicate that signer-dependent evaluation can substantially overestimate SLT capability. We recommend: (1) adopting signer-independent protocols to ensure generalisation to unseen signers; (2) restructuring datasets to include explicit signer-independent, sentence-disjoint splits for consistent benchmarking; and (3) reporting both signer-dependent and signer-independent results together with train-test sentence overlap to improve transparency and comparability.

Traditional light clients rely on validators committing to the entire blockchain state at every block via a state commitment such as a Merkle tree, allowing clients to verify facts using short proofs. However, maintaining large and ever-growing state trees imposes a significant burden on validators and lies on the critical path of block production. As a result, many modern high-throughput chains avoid this approach altogether. This work asks whether efficient inclusion proofs can be supported without requiring validators to maintain full state commitments. We present Guppy, a protocol that achieves this by having validators commit to just the state updates. An off-chain, untrusted service, secured by recursive Zero-Knowledge Proofs (ZKPs), then maintains a verifiable Merkle tree over the full state. This design keeps validator overhead negligible and does not increase the asymptotic complexity of block construction. Our design rests on two key technical ideas. First, a hash-chain commitment moves validator signature verification out of the ZK circuit, keeping the proving circuit efficient. Second, we design a parallel recursive proving pipeline that leverages cheap recursion in modern ZKPs to ensure latency grows only logarithmically with throughput. Our Plonky2-based implementation demonstrates that Guppy can maintain a Merkle tree of size 2^30 while processing thousands of updates per second, adding only 2-4 s of latency.

Graph modeling, a crucial task for representing complex relationships in graph-structured data, has achieved significant success in recent years. However, current graph modeling methods rely on traditional Graph Neural Networks and pre-training approaches to implicitly learn the underlying relational structure of graph data. Thus, these prior methods cannot capture the complex graph structure and correlations among inputs. In this paper, we introduce a novel Attention-based Normalizing Flow-based Approach\footnote{Our implementation and models will be released publicly for research reproducibility.} (ANFA or $\alpha$) that provides an explicit, interpretable, and tractable Graph Modeling ($\alpha$-Graph). In particular, we propose a new Unconditional Graph Normalizing Flow with an Invertible Attention Mechanism to capture the complex relational structure of graph data. To further enhance the expressiveness of the model, we introduce Conditional Graph Normalizing Flow with Learnable Queries that enables efficient modeling of correlations in graph-structured data. We show that our Conditional Graph Normalizing Flows behave similarly to Unconditional Graph Normalizing Flows, enhancing expressiveness while maintaining training stability and efficiency. Our experimental results on three benchmarks will illustrate the effectiveness and the state-of-the-art (SoTA) performance of the proposed $\alpha$-Graph method.

Fourier-Motzkin elimination computes an inequality description of the projection of a polyhedron onto a subset of its coordinates by eliminating one variable at a time. It is used in several areas of optimisation and computer science, and it is a standard way of obtaining the entropic constraints of a causal structure, where the marginalisation over the latent variables produces such a projection. Its limitation is the growth of the intermediate systems of inequalities, which can be doubly exponential in the number of eliminated variables even though the projection itself grows only as a single exponential. In practice the computational overload of the method therefore depends on two choices: how the redundant inequalities are removed after each step, and the order in which the variables are eliminated. We consider both. We first show, by an explicit example, that Imbert's redundancy test cannot be interleaved with redundancy removal by linear programming. We show that the two methods, however, can be combined soundly if the derivation records used by Imbert's test are re-initialised after every step at which linear programming is used. We then propose a rule for choosing the elimination order of the variables that gives a significant computational advantage, however, at the cost of increased resource usage. We demonstrate this advantage on some random polytopes, where the rule reduces the running time by factors of between 6 and 25 compared with the same elimination under a fixed order. For entropic descriptions of causal structures, with more than 250 inequalities and more than 100 variables to eliminate, our rule keeps the number of inequalities handled at each step one to two orders of magnitude lower than a fixed order.

Machine learning models are widely used in clinical applications, social media, law enforcement and critical infrastructure. Verifying whether their outputs are biased against disadvantaged groups or individuals is crucial to ensuring they are fair and allowing their use in such settings. A rigorous risk assessment of possible fairness violations requires quantifying the uncertainty associated with selecting and estimating such models. Yet, this is rarely done in the literature, which focuses on identifying a single model with a suitable trade-off between predictive accuracy and fairness. In this paper, we move beyond point estimation and discuss frequentist and Bayesian approaches to uncertainty quantification for fair machine learning, with practical examples and implications for simulated and real data.

Tabular Foundation Models (TFMs) have recently demonstrated strong predictive performance through in-context learning, but their deployment in high-throughput data streams remains challenging due to communication overhead and latency. We propose HINT, a hierarchical inference framework that combines edge-based retrieval with cloud-based TFM inference. A graph-based approximate nearest neighbor memory maintained over a sliding window provides local predictions and uncertainty estimates, allowing confident samples to be processed locally while uncertain instances are selectively offloaded, together with their retrieved context, to a cloud-hosted TFM. The framework exposes an offloading threshold and a neighborhood retrieval policy that can be varied to balance predictive performance and communication cost. Experiments show HINT consistently identifies favorable trade-offs.

Sparse Sinkhorn layers use a fixed support graph to restrict transport between tokens. How does this graph control gradient propagation through the scaling iterations. We develop a fixed-support calculus showing that each row-column cycle induces a row-stochastic operator on column-potential perturbations modulo constants. Its transpose propagates zero-mass reverse-mode cotangents. The finite-cycle operator uses two distinct half-step transport plans; at a balanced fixed point it reduces to a two-step walk determined by a single plan. We derive the accompanying score and marginal source terms and use Dobrushin contraction and minorization to bound homogeneous and source-driven tail cotangents. Our main result characterizes when support and marginals guarantee one-step contraction uniformly over finite scores: every feasible face of the transportation polytope must have pairwise two-hop column overlap. Otherwise, suitable score directions make the contraction coefficient arbitrarily close to one. We extend this analysis to ordered support schedules and derive certificates for partition heat-bath layers, coordinate sweeps, forced shared mass, and register-augmented supports. These results provide mathematical criteria for support design in differentiable transport layers, with guarantees restricted to the fixed-support quotient-gradient component.

We study the local errors of classical machine-learning surrogate models, which approximate the time evolution of the one-dimensional viscous Burgers equation. Four models are compared on the same prediction task, using the spatial grid values directly: radial basis function (RBF) kernel ridge regression (KRR), linear Ridge, ExtraTrees, and Random Forests. Across all four models, the one-step residual, defined here as the true value minus the predicted value at each grid point, forms clear curved branches near predicted maxima and minima. A more detailed analysis of KRR shows that these errors are much more strongly related to the second spatial derivative, which measures local curvature, than to the first spatial derivative. Near a smooth extremum, predicted value and curvature form a local two-branch fold. Under our local curvature-based model of the residual, this fold predicts a leading-order near-parabolic relation between predicted value and residual. This geometric result motivates a direct test of the Burgers advection (transport) and diffusion (smoothing) terms. For KRR and Ridge, regression tests on held-out trajectories, a control that breaks the spatial alignment of the diffusion term, and a spectral test of high-frequency content are consistent with insufficient viscous smoothing at moderate and high viscosity. In this case, the surrogate retains more small-scale structure than the true future state. The same physical explanation is much weaker for the tree models. Finally, a correction that uses only predicted quantities reduces both one-step error and error during recursive rollout, where each prediction is used as the next input.

We improve the best known lower bound on the integrality gap for weighted unit-demand multicommodity flow on trees from $1/4$ to $2/5$, improving on the long-standing bound of Chekuri, Mydlarz, and Shepherd~\cite{CMS}. We give the proof in two stages. First, a surprisingly simple packing lemma and an inductive coloring argument give an intermediate bound of $4/11$. We then refine the argument to obtain $2/5$.

We disprove Kenig's 1994 spectral-radius conjecture for the double-layer operator in every dimension $d\ge2$. For every $t>1/2$ sufficiently close to $1/2$, we construct a bounded simply connected Lipschitz domain with connected boundary such that the double-layer operator on $L^2$ with respect to the surface measure has $\pm \mathrm{i} t$ in its Fredholm essential spectrum. In particular, the essential spectral radius is greater than $1/2$. In the plane, the boundary may be chosen differentiable everywhere and smooth away from a single point. Our planar construction develops a refinement mechanism that produces unbounded growth of adjoint resolvent solutions with fixed forcing while keeping the graph slopes uniformly bounded. A closed recursion for the density and its image under an auxiliary singular integral reduces the growth estimate to a certified inequality for Hermitian $2\times2$ matrices. The certificate uses continuous estimates, exact rational arithmetic and rigorous interval enclosures. Localisation and rescaled graph insertions then produce a weakly null sequence of normalised approximate eigenvectors for the adjoint double-layer operator on one bounded boundary. A compact intertwining relation on Cartesian products of the planar domain with cubes gives the higher-dimensional examples without changing the spectral parameter.

Relativization is concerned with comparing computational models with black-box access to an oracle. For promise problems, black-box access is not canonical due to inputs outside of the promise being unconstrained. We study two semantics for such access. Under robust queries, a machine must correctly answer regardless of the completion of the problem,, while loose access requires that the internal choices of a machine do not change based on off-promise queries. Our first result separates the language and promise settings. Namely, we construct an oracle $O$ such that $\mathsf{P}^O = \mathsf{BQP}^O = \mathsf{AWPP}^O$, but $\mathsf{PromiseBQP}^O\not\subseteq\mathsf{PromiseP}^O_{\mathsf{/poly}}$. In particular, $\mathsf{BPP}^O = \mathsf{BQP}^O$, but $\mathsf{PromiseBQP}^O \neq \mathsf{PromiseBPP}^O$, showing that results for languages need not transfer to promises. Next, we use loose queries to strengthen the upper bound on the Quantum-Classical Polynomial Hierarchy from $\mathsf{P}^{\mathsf{PP}^{\mathsf{PP}}}$ to $\mathsf{QCPH} \subseteq \mathsf{BP\cdot PP} \subseteq \mathsf{PromiseBPP}^{\mathsf{PP}}$. The same proof also shows $\mathsf{PP}^\mathsf{PromiseBQP} = \mathsf{PP}$. Additionally, we show that $\mathsf{PromiseBQP}$, even when given quantum advice, is self-low under robust queries. Finally, we exhibit an obstruction to transferring language-level counting results to promise classes. Although $\mathsf{AWPP}$ and $\mathsf{APP}$ are low for $\mathsf{PP}$, a corresponding promise analogue would collapse the counting hierarchy as $\mathsf{GapP} \subseteq \mathsf{FP}^{\mathsf{PromiseAWPP}}$. This motivates the introduction of $\mathsf{PromisePostBQP^*}$, which restricts $\mathsf{PostBQP}$ to input-indepencent postselection. By showing that it is low for \PP, we obtain $\mathsf{PP}^{\mathsf{PromiseYQP^*}} = \mathsf{PP}$.

Language models increasingly perform empirical analyses end to end, yet existing evaluations assess the written explanation or whether generated code executes, not whether the executed workflow recovers the intended causal estimand. We introduce CausalVerify, an execution-grounded benchmark for end-to-end causal analysis that follows a model from research-context interpretation to estimand recovery. It scores this workflow at four distinct layers: method recognition, design specification, executable implementation, and estimand recovery. It combines 259 real-paper contexts, 100 fixed-seed synthetic scenarios with executable reference estimates, and 23 paper-twin pairs in which a model commits to a design before seeing the data and its executed analysis is scored against a canonical estimator on the same realised dataset. Execution is not correctness. Among 426 model-written workflows that run without error, 15.5% fail verification, and a keyword score of the effect direction stated in the text is a poor proxy for recovery. In a 23-pair, nine-model paired study, replacing a model's committed design with the reference design and its execution conventions raises joint recovery of the point estimate and standard error from 15.0% to 51.5%; yet 48.5% of eligible seeds still fail under the reference design. The direction replicates on six pairs built afterwards under a frozen construction protocol, although on the two newest pairs the gain is confined to models from the family that built the references. Design specification is consequential but not sufficient: a plausible method and runnable code do not guarantee recovery, and even supplying the reference design and its conventions leaves substantial downstream failure. CausalVerify evaluates the executed workflow rather than its surface plausibility, and every reported number is recomputed from frozen artifacts by a single script.

There is significant uncertainty about whether abstractions like beliefs or desires usefully describe the behavior of large language models (LLMs). In addition to the inherent scientific interest of this question, these latent quantities are often invoked to explain the behavior of LLMs to users or to define and evaluate harmful behaviors which are relative to intent. Nevertheless, we currently lack a means to systematically test whether concepts like "belief" are well-applied to LLMs, and hence whether they are likely to be fruitful ingredients of attempts to align models with human interests. We propose an approach for empirically studying such questions, asking whether a single latent variable inferred from the LLMs' outputs – interpreted as a degree of belief – allows an observer to make interpretable predictions of how the LLMs' will respond to new prompts. We find that highly capable models are usefully described as holding beliefs and that, generally, the predictability of model outputs based on an inferred latent belief tracks overall trends in model capability. Building on these findings, we provide empirical strategies to study how beliefs in LLMs can be measured, the extent to which LLMs comply with instructed decision rules or payoffs, and how beliefs evolve within individual instances of an LLM over the course of reasoning.

In this paper, we propose ReactVAU, a Slow-Fast Decoupled Framework for real-time streaming Video Anomaly Understanding (VAU). Existing VAU methods rely on offline inference with global temporal sampling, which violates causality and prevents deployment in live surveillance streams. Conversely, general streaming video models satisfy causal access but dilute rare transient anomalies during memory compression and often invoke heavyweight MLLMs uniformly over long normal intervals. React VAU addresses this gap with three synergistic components: a lightweight Fast Detection Module based on Spatial Grid Folding (SGF) for continuous anomaly filtering; an Anomaly-Aware Persistent Memory (AAPM) that protects critical visual cues from temporal decay; and a heavyweight Slow Reasoning Module that remains dormant during normal streams and is awakened only by suspicious events for semantic verification and causal description. Extensive experiments on multiple benchmarks demonstrate that ReactVAU operates under strict streaming constraints while simultaneously achieving competitive performance in both anomaly detection and causal reasoning, alongside significantly enhanced computational efficiency by minimizing heavyweight MLLM invocations. Project page is available at https://huiyuiui.github.io/ReactVAU/

We address the problem of discovering repeated elements from a single image. In contrast to existing approaches that depend on large annotated datasets, curated multi-image collections, or object segmentation masks, we show that a single image can suffice to learn a meaningful object model in a completely bottom-up fashion, without any prior knowledge beyond a coarse scale prior. Our method learns a tunable image-space prototype of the repeated elements through a reconstruction objective, enabling the model to identify and synthesize consistent object instances within the same image. Experiments on 116 real images from the FSC-147 dataset demonstrate that our method successfully learns coherent element models and captures intra-category variation on challenging images. Qualitative results reveal superior reconstructions and interpretable decompositions compared to classical decomposition, joint alignment, and 3D object modeling methods, while maintaining a simple 2D formulation. These results suggest that meaningful object discovery can emerge from single image learning alone.

We propose a new fractional threshold-dynamics algorithm for subdiffusive interface motion and apply it to multiphase grain-growth simulations. The method combines the L1 discretization of the Caputo time derivative with a Helmholtz-resolvent thresholding step. At each time level, the L1 convolution produces a convex history average of the previously thresholded states; this history-dependent field is then propagated by a Helmholtz diffusion solve and projected back to pure phases by thresholding. For $\alpha=1$, the method reduces to a Helmholtz-resolvent analogue of the Merriman-Bence-Osher scheme, whereas for $0<\alpha<1$ it introduces a subdiffusive temporal memory. We establish a variational characterization, maximum and comparison principles, and stability for the continuous Helmholtz diffusion step. We also derive a memory-tail estimate and a finite-dimensional pinning criterion under an explicit resolvent-contraction hypothesis. The equal-tension multiphase scheme is implemented on periodic grids. The numerical experiments investigate the joint influence of the L1 history and the smoothing length. The fixed-step runs show faster coarsening for smaller fractional orders in the tested regime, while a comparison at fixed smoothing length examines the influence of the L1 history on successive updates.

Structured visual reasoning, such as image puzzles, demands fine-grained visual perception, which current Vision Language Models (VLMs) often lack. VLMs built on CLIP-based ViT backbones tend to trade fine-grained visual details for high-level semantics. We show that this deficit propagates downstream, affecting various vision and language tasks. To address this, we fuse DINOv3 and CleanDIFT representations into a perception encoder (DiffusedDINO), and align it with RoBERTa-L, yielding a text-aligned model (TDDN) that preserves this perceptual advantage: with frozen backbones and only ~590K alignment pairs, TDDN surpasses CLIP ViT-L/14 in image-text retrieval tasks: text-to-image on Flickr30K and both directions on MS-COCO-2014. TDDN more than triples CLIP's dense-prediction accuracy (ADE20K 5.20 -> 18.11 mIoU, COCO-Stuff 7.35 -> 24.44), despite CLIP's massive training corpus. TDDN leads on segmentation benchmarks among general-purpose contrastive encoders, including SigLIP 2. We further introduce PuzzlePerception, a segmentation and visual question answering dataset that probes fine-grained spatial understanding, on which TDDN doubles CLIP's segmentation accuracy (11.04 -> 22.51 mIoU). Our code and data are available at https://github.com/adityaSomak/TDDN.

As geopolitical tensions, energy crises, and energy-intensive AI infrastructure intensify concerns about demand, affordability, and resilience, communities increasingly encounter these challenges through everyday energy practices, particularly winter heating. Against this background, the doctoral exposé, "Designing Human-HVAC Interaction for Healthy, Affordable, and Sustainable Heating in Smart Homes", is structured around four chapters. First, a multidisciplinary literature review defines and positions Human-HVAC Interaction, focusing on heating in smart homes. Second, longitudinal living lab studies with design probes examine everyday heating practices, thermal comfort, and indoor environmental quality, with attention to thermally vulnerable groups such as older adults, pregnant or menopausal women, parents with infants, and people affected by allergies or airborne pollutants. Third, a VR-based smart home demonstrator explores how heating and IEQ scenarios can be prototyped and evaluated as a virtual living lab, while critically examining the limits of representing bodily indoor climate conditions through VR. Fourth, follow-up design studies examine how VR-based insights can be translated into physical-digital prototypes that combine digital fabrication, distributed environmental sensing, and diverse interface forms for critical heating and IEQ contexts. The thesis aims to contribute a design-oriented understanding of Human-HVAC Interaction by building from a multidisciplinary literature review to empirical living lab and co-design studies, VR-based prototyping, and physical system development, examining how smart home users make sense of, negotiate, and respond to smart HVAC system.

High-quality demonstration data is becoming a central bottleneck for training general-purpose humanoid robots. While recent humanoid teleoperation systems have made substantial progress in retargeting human motion to robot motion, long-horizon loco-manipulation requires another capability: operators must maintain task-relevant spatial awareness over time, e.g., object locations, surrounding environments, the robot's pose. We call the extent of this awareness the operator's perceptual horizon. However, existing methods often shorten this: narrow views miss peripheral events, robot-mounted cameras become unstable during locomotion, and coupled head-view control makes looking around interfere with robot motion. We present SPOT, a Spatial Perception-Oriented VR Teleoperation system for collecting long-horizon humanoid demonstration data by providing extended perceptual horizon. SPOT combines a robot-mounted binocular fisheye camera, a wide-field stereoscopic display, viewpoint-decoupled free-looking, and visual stabilization to provide a robot-centric view that is wide, stable, and actively inspectable. Unlike conventional egocentric interfaces, SPOT decouples visual exploration from robot actuation: the egocentric stereo observation is rendered on a virtual hemisphere around the operator, so natural head rotations change where the operator looks within the wide-field view rather than commanding the robot head, camera, or torso. We evaluate SPOT on perception-critical humanoid data-collection tasks spanning drop recovery, peripheral retrieval, large-workspace bimanual manipulation, fine alignment, and dynamic interaction. SPOT improves efficiency, accuracy, and recovery speed, demonstrating its effectiveness for user-friendly and scalable long-horizon humanoid data collection.

Deep learning models have been increasingly applied to Time Series Forecasting (TSF) in recent years. Transformer-based and MLP-based models have both been used effectively on many real-world TSF regression benchmarks, and there is ongoing debate as to which family of methods is best. While these benchmarks have drawn much attention, it is also worth noting that many current datasets and methods assume approximate periodicity in the time series. In this work, we focus on a new TSF task without periodicity: anticipating falls during humanoid locomotion, on the basis of egocentric vision and proprioception. When the locomotion trajectories are sufficiently diverse, periodicity is violated. We contribute two new benchmark datasets (one from simulation, one from real hardware), showing that periodicity is violated and recent deep TSF methods struggle on these benchmarks. We also propose a novel deep learning architecture that exploits both endogenous and exogenous variables and a training process that rigorously enforces i.i.d sampling of training examples. Our results show statistically significant improvement over prior art in multiple experimental conditions, by 12.73% or more on the real data and 10.40% or more on the simulation data. Code and datasets will be available upon acceptance.

The placement of microservices in the compute continuum plays a vital role in delivering services that comply with customers' needs, such as reduced latency, storage requirements, quality of service and availability. To achieve customers' needs in the geographically dispersed architecture of the compute continuum, Service Level Objectives (SLOs) have been largely used in decision-making to place microservices. However, because low-level SLOs increase the barrier to entry for continuum users, placement decisions based on high-level business vocabulary are required if the compute continuum is to be adopted at scale. This paper proposes an architecture for microservices placement decisions in the computing continuum utilizing high-level user intents described in natural language as input. The approach utilizes Generative Artificial Intelligence to translate the intents to low-level SLOs, which are used along with the infrastructure description to decide where different microservices that compose an application must be deployed so that SLOs are met. We implement and evaluate a prototype of the architecture to demonstrate the approach's feasibility.

We characterize the degree of freedom (DoF) of a point-to-point blockwise memoryless channel without prior channel state information (CSI), with a fixed number $K$ of propagation paths, where the transmitter (Tx) and the receiver (Rx) are equipped with nonuniform linear arrays (NULAs) of $N_t$ and $N_r$ antennas, respectively. The positions of array elements are fixed, known, pairwise distinct, and need not be equally spaced. The uniform linear array (ULA) is a special case. In each block of length $T$, the continuous angles of arrival (AoAs), angles of departure (AoDs), and independent complex Gaussian path gains are redrawn. Both Tx and Rx know the state distributions but are not given the current realizations before transmission. The receiver may estimate the channel from reference signals or decode without explicit channel estimation, with reference symbols counted in $T$ and their energy counted against the power constraint. Under the aforementioned model, we show that the DoF is $1-\frac{1}{T}$ for $K=1$, and $K(1-\frac{3}{2T})$ for $K \geq 2$, when $N_r\ge K+1$, $N_t\ge\max\{K,2\}$, and $T\ge K$. The analytical results are further demonstrated by their applications to the DoF tradeoff analysis in integrated sensing and communication (ISAC). For more general array structures, an achievability result is established, while the converse remains open in general.

We present FrogNano, a 4B coding agent designed to tackle software engineering (SWE) tasks efficiently and effectively, even under resource-constrained environments. It is post-trained exclusively via RL on around 1,500 SWE environments with synthetic tasks. A key ingredient for improving performance is an online task synthesis pipeline that creates tasks calibrated to the frontier of learnability for the current checkpoint. This report provides evidence that competitive small coding agents can be trained with synthetic tasks alone, without traditional distillation from larger models, and that generating tasks at the learnability frontier of the current agent is important. We report details on the training methodology, evaluations across diverse environments, and in-depth analyses, serving as a foundation for our ongoing exploration of lightweight yet capable coding agents that can run on minimal hardware.

Shortcut learning denotes the widespread situation in which a classifier exploits spurious correlations rather than diagnostic features. Existing mitigation strategies mostly aim to learn shortcut-invariant representations; their empirical success is limited and they cannot be applied to classifiers using frozen foundation model encoders. We propose to reframe shortcut learning as fundamentally a calibration problem: unconstrained learning implicitly calibrates each shortcut group to its training set disease prevalence, rendering the resulting classifier necessarily over-confident in one group and under-confident in the other. Building on this insight, we prevalence-equalize calibration between shortcut groups through two encoder-agnostic methods, an in-processing regularizer and a post-hoc prevalence-equalized recalibration step. Across chest-drain-pneumothorax benchmarks on CheXpert and SIIM-ACR, spanning fine-tuned CNNs and frozen foundation-model backbones, both methods substantially outperform all baselines. Post-hoc recalibration of a standard ERM-trained DenseNet raises misaligned-group AUROC from 0.23 to 0.73, indicating that shortcut reliance degrades the classification head rather than the underlying representation. Besides two new state-of-the-art shortcut mitigation approaches, our findings more fundamentally connect shortcut learning to calibration theory and algorithmic fairness.

Human-LLM co-writing is increasingly used for open-ended text generation, but much prior work focuses on final outputs rather than the interactional dynamics through which stories are produced. We study turn-based collaborative storytelling across three matched conditions: Human-Human (HH), Human-LLM (HA), and LLM-LLM (AA). Using a shared storytelling paradigm, we measure how agents align, introduce novel material, and influence narrative development through turn-level measures of valence adaptation, semantic novelty, transience, and resonance. Our results show that HA co-writing is not intermediate between HH and AA collaboration. Instead, it displays a distinctive asymmetry where humans tend to introduce more novel and persistent narrative material, while LLMs tend to elaborate and stabilize the existing context. These findings suggest that, in this setting, LLMs function less as human co-authors and more as adaptive narrative amplifiers that reshape how agency is distributed in collaborative writing.