Computer Science (arXiv)

Data analyses are usually designed to identify some property of the population from which the data are drawn, generalizing beyond the specific data sample. For this reason, data analyses are often designed to produce a low generalization error, so that the result of an analysis on sample data does not differ too much from the result one would achieve over the entire population. An adaptive data analysis can be seen as a process composed of multiple queries interrogating some data, where the choice of which query to run next may rely on the results of previous queries. The generalization error of each individual query can be controlled using well-established statistical techniques. However, when queries are arbitrarily composed, errors can propagate through the chain of queries and lead to high generalization error. To address this issue, several techniques guarantee bounds not only on single queries but also on composed analyses. The choice of technique often depends on the chain of queries that an adaptive data analysis can generate. In this work, we consider adaptive data analyses implemented as while-like programs and design a program analysis to help identify which technique to use to control their generalization errors. More specifically, we formalize the intuitive notion of adaptivity as a quantitative property of programs. Based on this definition, we design a program analysis for soundly approximating this quantity. The analysis represents the data analysis as a weighted dependency graph, where weights upper-bound how often variables can be reached, and uses a path-search strategy to upper-bound adaptivity. We implement our program analysis and show that it can analyze several concrete data analyses with different adaptivity structures.

World action models jointly predict future visual observations and actions, whereas existing tactile-aware variants typically represent future touch as an image or latent stream without modeling the physical dependencies that organize tactile states hierarchically. We present HiTac-WAM, a hierarchical tactile world action model that forecasts a sequence of future tactile states for each candidate action chunk before execution. The forecast factorizes into contact state, a 3D deformation field, and slip risk, organized as a directed hierarchy in which each downstream stage is conditioned on stop-gradient signals from preceding stages. A directed attention mask allows tactile queries to attend to the video-action context of each candidate while preventing video and action queries from attending to tactile tokens. For planning, HiTac-WAM ranks candidate action chunks using tactile forecasts and task-progress estimates. For execution, the selected tactile forecast is retained as a reference; persistent discrepancies between predicted and observed tactile states trigger corrective replanning. HiTac-WAM achieves a mean contact F1 of 0.921; under matched training budgets, the directed hierarchy reduces 3D displacement L2 error by 17.6% relative to the deformation-only predictor and improves slip AUPRC by 60.4% relative to the slip-only predictor. Across chip grasping, blackboard erasing, and USB insertion, selection guided by the hierarchical forecasts increases the average real-robot success rate from 31.1% to 61.1%, while the full system attains 72.2%.

The Boolean connectivity problem asks whether the set of satisfying assignments of a given Boolean formula forms a connected subgraph in the $n$-dimensional hypercube. This problem is known to be $\mathsf{coNP}$-complete, even when restricted to $k$-Horn formulas for $k \geq 3$, as shown by Makino, Tamaki, and Yamamoto. In this paper, we further investigate the computational complexity of {\sc Conn $k$-Horn}, the Boolean connectivity problem for $k$-Horn formulas. We provide algorithmic and hardness results for {\sc Conn $k$-Horn}. On the algorithmic side, we first present an exact exponential-time algorithm for arbitrary $k$ without any structural restrictions. Our algorithm builds on the deterministic PPZ algorithm proposed by Paturi, Pudlák, and Zane. It runs in $O^*(2^{(1 - 1/2k)n})$ time and polynomial space, achieving an exponential improvement over the previously known algorithm for the Boolean connectivity problem of $k$-CNF formulas, shown by Makino, Tamaki, and Yamamoto. We next give two polynomial-time algorithms for arbitrary $k$ under the following two restrictions: (i) each variable appears at most twice, and (ii) each clause has length exactly $k$ and each variable appears at most $k$ times. On the hardness side, we prove that {\sc Conn $3$-Horn} remains $\mathsf{coNP}$-complete even when each variable appears exactly three times.

Drafting a new Magic: The Gathering expansion begins before any pick from it has been observed: the complete card list is public, but the draft logs that supervised pick models train on do not yet exist. We study this day-zero regime directly. DraftFM is a discrete-choice policy that scores exactly the cards available in the current pack, conditioned on the drafted pool and the state of the draft. Every card enters as a frozen 775-dimensional function of its public card record, structured features and a fixed text embedding, with no card identities, set identities, or usage statistics anywhere in the model, so an unseen card is scored by the same machinery as a familiar one. A 1.6-million-parameter network fitted on 149 million human picks from 29 expansions predicts held-out picks in three expansions withheld in their entirety, reaching 50.8%, 60.4%, and 56.7% top-1 agreement, where uniform chance at the opening pick is about 7%. Refitted on all 32 observed expansions, the same architecture produced a card ranking for the then-unreleased set The Hobbit, sealed with its complete cryptographic provenance and published roughly 36 hours before the set became draftable on MTG Arena. The sealed ranking agrees with six independent expert reviewers roughly as much as those reviewers agree with one another. Evaluation against realized outcomes is committed to a follow-on note, whatever it shows.

While text-to-3D generation has advanced rapidly, achieving high geometric fidelity at low inference cost remains challenging. Existing text-to-3D methods either decode discrete shape tokens autoregressively or iteratively refine global 3D representations with diffusion or flow models. However, autoregressive decoding is sequential and cannot revise errors, whereas diffusion and flow-matching models repeatedly process the full representation, making high-quality generation increasingly expensive. In this paper, we propose Block3D, a block-wise diffusion framework that partitions the discrete shape-token sequence into contiguous blocks, generates the blocks autoregressively, and jointly denoises all tokens within the current block. To alleviate error accumulation, we introduce confidence-guided intra-block correction, which revises low-confidence tokens before each block is finalized. On a held-out set from TRELLIS-500K, Block3D reduces mean end-to-end generation time from 25.71 seconds to 4.99 seconds, achieving a $5.15\times$ speedup over the fine-tuned autoregressive baseline without sacrificing geometric fidelity.

Persistent memory can personalize an LLM agent, but an incorrect durable update can silently distort future behavior. We study the memory-clarification boundary: whether interaction-derived information should be persisted, used only in the current context, re-verified, or clarified with the user. MCB contains 140 primary scenarios, split into 70 development and 70 held-out items, plus a separate 70-item contrast set. It evaluates both action labels and structured tool-call selection. Two non-authors independently label the 70 held-out primary and 70 contrast items (97.1% agreement, Cohen's kappa = 0.962); a blind third resolves four disagreements, replacing eight author labels by non-author majority. Across Claude and Qwen, models verify changing facts more reliably than they ask users to resolve ambiguity. Bare Qwen asks on 0/12 clarification items while verifying 12/18 freshness items. Few-shot prompting raises accuracy from 0.557 to 0.771 (paired delta = +0.214, Holm-adjusted exact McNemar p_H = 0.002), yet clarification recall remains 0.333. The policy prompt reduces erroneous persistence from 0.243 to 0.100 (p_H = 0.038), although its accuracy gain is not significant. Label-tool agreement is 57% for each Claude model and 23% for Qwen; Qwen accuracy falls from 0.557 to 0.343 (p_H = 0.047). Memory evaluation must test both stated decisions and tool-call choices.

Financial AI systems often train information extractors on one textual register and deploy them across filings, news, and user-generated content, while standard F1 scores do not indicate which predictions remain safe to automate when the input distribution changes. We study confidence estimation and selective prediction for financial named entity recognition (NER) on a three-tier stress test spanning SEC filings, financial news, and general-topic social media as an extreme out-of-domain condition. We evaluate a BERT tagger and LoRA-tuned Qwen2.5-0.5B/1.5B models using five inference-time confidence signals, three training seeds, and bootstrap intervals. Confidence rankings themselves change under distribution shift: whole-output probability is the strongest in-domain error detector but deteriorates out of domain, whereas entity-span probability and self-consistency are more robust; self-consistency is also better calibrated without post-hoc fitting. Abstention reduces sentence error from 34.3% to below 2% on the highest-confidence 40% of in-domain inputs and remains useful on financial news, but recovers no usefully large clean subset under the extreme social-media shift. These results motivate a staged deployment strategy that detects severe distribution shift upstream before applying prediction-level confidence gating.

Autonomous vehicles offload latency-sensitive perception tasks to nearby mobile edge computing (MEC) servers, where a missed safety-critical task is unsafe rather than merely degraded. Large language models (LLMs) are increasingly proposed as adaptive, explainable schedulers, yet evidence of when they help is scarce. We study deadline-aware, mixed-criticality scheduling on heterogeneous MEC servers, where time-critical (TC) tasks must be protected at a controlled cost to best-effort traffic, and ask whether a multi-agent LLM control layer improves on a strong heuristic. We answer in two steps. First we build the heuristic: a windowed contract-net auction that orders each admission window time-critical-first by earliest deadline and places tasks by earliest-finish-time. Across 60 instances on three topologies and 15 baselines under an identical online constraint, it attains a TC completion rate of 0.902, above every baseline (Holm-corrected p < 0.001; best baseline 0.838) and at 0.87 of a CP-SAT upper bound. Second, we add the LLM control plane. A controlled decomposition traces the scheduler's advantage to two ordinary factors, the batching horizon and time-critical-first ordering; the auction, the per-window LLM policy, and online adaptation add nothing while the load is stationary, where the heuristic is already near-optimal. Under a mid-run surge of safety-critical tasks the picture changes, and the LLM control plane gains significantly over both the static heuristic and the bandit. LLM orchestration therefore earns its cost only when non-stationarity opens headroom a fixed policy cannot use. We report control-plane latency and rationale, and release all code and seeded instances.

Streaming autoregressive diffusion models enable real-time, long-horizon video generation, but their training objectives optimize local frame prediction rather than the geometry and dynamics of a coherent world: long rollouts accumulate geometric drift and degrade into static or unnatural motion. Recent bidirectional approaches address this problem using rewards signals built upon 3D Gaussian-Splatting reconstruction. However, a single rigid 3d reconstruction cannot model a dynamic scene, so this critic penalizes genuine object motion as reconstruction error and is maximized by freezing the video. This shortcut is especially detrimental in the AR setting, where each chunk can propagate an already-static configuration. In this work, we propose Stream4D, which replaces the static critic with a feed-forward 4D reconstruction reward that explicitly models scene dynamics, allowing coherent motion to receive high consistency rewards. To further guide motion magnitude and quality, we add a motion prior that rewards natural scene-flow magnitude while penalizing jitter and non-rigid artifacts. Our final recipe combines these two terms with a lightweight perceptual anchor. Across various autoregressive video backbones and various generation horizons, Stream4D improves 4D reconstruction quality, preserves motion more effectively, and achieves higher human-aligned preference. Project page: https://banyuanhao.github.io/Stream4D/

Deploying broadband Mission-Critical Push-To-Talk (MCPTT) services over shared commercial infrastructures introduces resource contention during multi-agency responses in mass-crowd events. This study evaluates cross-carrier interoperability and standard versus prioritized quality of service (QoS) frameworks under real-world saturation constraints. We design an empirical multi-carrier field experiment utilizing twelve identical smartphones deployed across multiple physical sectors inside Texas A&M University's Kyle Field during a football game with 105,000+ attendees. Automated voice calls were monitored using Perceptual Objective Listening Quality Analysis (POLQA), packet delivery metrics, and connection rates. The results reveal that voice path failure is isolated to network infrastructure bottlenecks rather than device hardware limitations. Specifically, we identify a sharp, non-linear network failure model where transport-layer jitter exceeding a critical threshold de-jitter buffer underflows, causing structural audio degradation. Priority-managed channels effectively bypass this congestion. This study helps establish an operational insight for emergency planners to mandate network infrastructure, end-to-end network slicing and dedicated resource provisioning capable of keeping transport-layer jitter below the critical failure boundary.

Vision--language models can identify the correct referent while returning an imprecise bounding box. We study whether a frozen direct-answer model can use its own prediction to allocate one additional localized observation without accessing target annotations at inference. Label-free precision refinement (LFPR) routes predicted-small regions to a higher-resolution pass, re-grounds the expression inside a context crop, admits a candidate only under fixed geometric guards, and returns a fixed coordinate-wise midpoint. We report results across three evidence tiers. On 31,921 retrospective Ref-L4 expressions, LFPR raises mAcc$_{0.5:0.95}$ from 72.947\% to 76.013\% (Acc@0.5 88.531\%$\to$89.725\%, Acc@0.9 55.788\%$\to$61.142\%). A frozen transfer to 30,969 RefCOCO/RefCOCO+/RefCOCOg expressions improves every dataset at Acc@0.5, mAcc, and mean IoU (pooled mAcc $+0.645$, Acc@0.5 $+0.817$), while Acc@0.9 is unchanged overall: routing alone gains $+1.162$ points there, but crop, guards, and fusion give back $-1.192$, offsetting rather than showing no strict-IoU effect. A prospective, image-disjoint Flickr30K Entities evaluation improves every endpoint (mAcc $+0.973$, Acc@0.9 $+1.022$), more strongly under a single-box variant (mAcc $+2.575$, Acc@0.9 $+3.689$). The same operator applied to two released grounding specialists improves every endpoint (Acc@0.9 $+1.569$/$+6.716$ for EGM-4B/8B) at roughly twice the latency, composing with specialist training rather than replacing it. A genuine unguarded control (guard removed from the same candidates) underperforms the incumbent on every metric, showing the guard is load-bearing. Together, these results show that referent selection and boundary precision are partially separable, with different components moving opposing regions of the IoU curve -- behavior a single threshold cannot reveal.

Accelerating the electrification of thermal energy requires vapor-compression heat pumps capable of dynamic, grid-responsive operation. However, equipment engineering remains fragmented across static rating-point selection, stiff multi-phase transient simulation, and gradient-based optimal control. Here, we present an end-to-end differentiable, finite-volume vapor-compression framework implemented natively in JAX that automates machine sizing directly from stated thermal duties and unifies dynamic simulation with predictive control under a single compiled residual $\dot{y}={f}(t,{y},{u})$. Thermodynamic evaluations bypass runtime root-finding via bilinear $(p,h)$ manifolds pre-flashed from Helmholtz equations of state, enabling analytical forward-mode automatic differentiation. Mass conservation across multi-phase coils is strictly preserved by incorporating both $(\partialρ/\partial p)_h$ and $(\partialρ/\partial h)_p$ partial derivatives into the dynamic pressure differential equation. The sizer directly inverts compressor displacement, electronic expansion valve area, and heat-exchanger tube counts via four-point cycle synthesis and $\varepsilon$-NTU matching using the identical polytropic compressor map. Crucially, the compiled physics kernel is shared symmetrically between $L$-stable TR-BDF2 stiff integration and implicit-Euler Model Predictive Control (MPC), eliminating plant-controller surrogate mismatch. Validated against open-access experimental benchmarks without parameter fitting, the framework predicts cooling capacity with $7.37\%$ MAPE across 16 mini-split operational runs and bounds on-period cooling error within $1.19\%$--$1.62\%$ on utility-scale Hardware-in-the-Loop traces. This work provides an open-source, differentiable foundation for automated machine synthesis, dynamic grid orchestration, and gradient-based hardware-control co-design.

Large Language Models (LLMs) are increasingly embedded into applications, allowing users to complete tasks either through direct manipulation or by delegating actions to conversational agents. However, little is known about how users balance these modalities when both are available. We present a web-based content management system augmented with an LLM agent through the Model Context Protocol (MCP), enabling users to perform CRUD tasks through a graphical interface, a conversational agent, or both. We conducted a between-subjects study (N=73) comparing three interaction modes: Traditional-Only, AI-First, and Hybrid. Across sixteen scenarios, we analyzed task completion time, interaction logs, and delegation behavior. AI-assisted interaction significantly reduced clicks, page navigations, and scrolling indicating lower interaction effort. Surprisingly, these reductions did not translate into faster task completion, as task duration did not differ significantly across conditions. We also found no significant relationship between CRUD operation type and delegation, suggesting that users did not systematically avoid delegating higher-risk actions. Instead, delegation varied far more between participants than between tasks, with individual differences accounting for roughly half the variance in assistant use (ICC = .50). Our findings suggest that the primary benefit of human--agent interfaces may be reducing interaction effort rather than improving speed, and that delegation reflects who the user is more than what the task demands.

Weak arcs are point sets in PG$(n-1,q)$ meeting every general hyperplane (those are the hyperplanes not going through one of the points given by the standard basis vectors) in at most $n-1$ points. In this paper, we study weak arcs together with balanced variants which are contained on the sides of the fundamental simplex. We give an upper bound on the size of weak arcs, characterise the largest balanced quasi-arcs in the plane and construct large balanced quasi-arcs in PG$(3, q)$. We then use these configurations to build point sets for the random-access problem in DNA-based storage. The constructions are explicit, work over small fields, and attain recovery expectations matching the best known asymptotic bounds.

This paper addresses the issue of the significant labor required to test interview dialogue systems. While interview dialogue systems are expected to be useful in various scenarios, like other dialogue systems, testing them with human users requires significant effort and cost. Therefore, testing with user simulators can be beneficial. Since most conventional user simulators have been primarily designed for training task-oriented dialogue systems, little attention has been paid to the personas of the simulated users. During development, testing interview dialogue systems requires simulating a wide range of user behaviors, but manually creating a large number of personas is labor-intensive. We propose a method that automatically generates personas for user simulators using a large language model. Furthermore, by assigning personality traits related to communication styles when generating personas, we aim to increase the diversity of communication styles in the user simulator. Experimental results show that the proposed method enables the user simulator to generate utterances with greater variation.

Autonomous conversational agents and generative-AI features are being added to online dating platforms faster than public evidence about user attitudes can accumulate, and the scarcest evidence concerns the receiving side: how people react when the profiles, messages, or conversation partners they encounter are machine-generated. We release two anonymized survey datasets collected from active users of Fledge.Love, a dating platform serving an international user base. The first (N = 2,617; Russian and English forms) measures receptivity to autonomous conversational agents with a seven-item battery that separates the principal role (deploying one's own agent) from the counterpart role (encountering someone else's), plus six ordinal covariates and two auxiliary items. The second (N = 2,894) measures interest in three passive generative-AI features. The release includes model-derived scores for 2,499 complete cases, a bilingual codebook, a documented anonymization pipeline with a k-anonymity audit, executable analysis notebooks, and canonical outputs, supporting reuse in human-AI communication, recommender-systems, and cross-cultural technology-acceptance research.

Functionalities of ferroelectric materials are governed by the spatial organization and coupling of polarization, strain, lattice rotation, and structural order accessible via atomically resolved scanning transmission electron microscopy (STEM) images. Quantitative interpretation of atomic-resolution STEM data has conventionally relied on locating atomic columns and converting their fitted coordinates into local structural descriptors. Here, we develop a field-based approach in which atomic-resolution images are represented by spatially varying latent Bragg fields, whose amplitudes and phases provide continuous maps of crystalline order, lattice displacement, strain, rotation, and mode-specific residual structure. The observed atomically resolved images are decoded from the latent fields. We apply this framework to image series of Sm-substituted BiFeO3 spanning 0-20% Sm and crossing the composition-driven boundary between the R3c ferroelectric phase and the orthorhombic, nonpolar Pnma phase. Conventional atom-resolved parameterization is used as an independent validation, showing that reconstructed Bragg amplitude tracks local atomic-column intensity and that field-derived shear reproduces unit-cell angular distortions obtained from atom fitting. The combined analysis reveals a systematic evolution from extended ferroelectric domains at low Sm concentration, through the appearance and growth of localized regions with period-doubled Pnma order at intermediate compositions, to a connected Pnma-dominated state at high Sm content. The period-doubled order is accompanied by enhanced shear and lattice rotation and by progressive reorganization of the ferroelectric domain structure. These results establish latent-field reconstruction as a physically interpretable complement to atom finding and provide a unified framework for resolving composition-driven phase evolution in ferroic materials.

Training fast generators on new domains with limited data remains challenging for two reasons. First, adapting a pretrained diffusion or flow model to a new domain leaves its costly multi-step sampling unaddressed, and existing acceleration methods are tied to the source parameterization--$ε$, $x$, $v$, or $u$--leaving heterogeneous pretrained models with no common acceleration target. Second, while adversarial refinement is proven effective for few-step quality, it is formulated only for instantaneous-velocity flows, not for the finite-interval average velocities that MeanFlow (MF) models predict. We address both problems. We propose MeanFlow-Transfer, which maps heterogeneous source outputs into a shared velocity representation, uses it to initialize an MF generator from the source weights, and optimizes an MF objective on the target domain. This unifies adaptation and acceleration in a single training loop across a broad range of pretrained models. We then introduce Continuous Adversarial MeanFlow, a post-training stage that extends continuous adversarial flow models from instantaneous velocities to MF's finite-interval average velocities. CAMF contrasts changes in a learned potential between real and predicted interval endpoints, recovering fine detail that MF regression averages away, and reduces to the instantaneous criterion in the vanishing-interval limit. Adapting four ImageNet-based source models--DiT ($ε$), SiT ($v$), JiT ($x$), iMF ($u$)--to five target domains, MF-T with CAMF matches or exceeds the fine-tuned teacher in FID and FDD at up to $125\times$ fewer Neural Function Evaluations (NFEs), while CAMF improves MF-T's few-step FID by $29\%$ on average.

Computing shortest paths in large graphs is, and remains, a fundamental and practically motivated problem. While many algorithms were proposed to calculate shortest path between pairs of vertices efficiently, many of them (index-based methods) require substantial preprocessing, while others (traversal-based methods) have higher time complexity. In this paper, we propose and analyze Pruned Bidirectional Search (PBS), a simple sublinear approximation algorithm for power-law graphs with parameter $β\in[2,3)$: our algorithm does not require any preprocessing, yet exhibits performance comparable to light index-based algorithms (of linear or sublinear index size): that is, PBS runs in time $O(n^{(1-1/\log\log n)/2})$ and, with high probability, returns a path with length within $\frac{41}{32}$ of the shortest path. Moreover, if one does allow a $n^{Θ(2-1/\log\log n)}$-time preprocessing step, its query time improves to $n^{Θ(1/\log\log n})$. We complement our theoretical results by experiments on both real-world and synthetic power-law graphs, which show that PBS is typically $1.84\times$-$7.76\times$ times faster than existing alternatives, while achieving an approximation ratio at most 1.05.

This paper presents a guide path-free multimodal trajectory planning framework for autonomous surface vehicles operating in dynamic environments. The proposed method integrates model predictive control (MPC) with a turning circle-based control barrier function (TC-CBF). Unlike conventional Euclidean distance-based CBFs (ED-CBFs), which evaluate safety solely based on proximity, the TC-CBF accounts for the nonholonomic motion and finite turning capability of a surface vehicle. Its geometric formulation identifies feasible avoidance regions according to the vehicle's turning circles and generates distinct left- and right-turning avoidance modes. These modes allow the optimization solver to explore and select topologically different trajectories without relying on globally planned guide paths, as required by many conventional multimodal planning approaches. By embedding the avoidance direction directly into the safety constraint, the proposed framework alleviates the local-minimum and deadlock problems of single-mode MPC while maintaining computational efficiency. Extensive simulations involving multiple moving vessels demonstrate that the proposed method achieves higher success rates, fewer safety violations, and smaller residual violations than single-mode baselines across all tested traffic densities.

Learning-based global point cloud registration has achieved remarkable progress, yet its reliance on geometric representations makes existing methods sensitive to variations in point density, scan pattern, viewpoint, and sensor characteristics. We propose CVSD-Reg, a robust global LiDAR registration framework that distills visual semantic priors from a vision foundation model into LiDAR representations. In Stage 1, a Point Transformer V3 student learns from a frozen DINOv2 teacher through contrastive distillation and spherical-manifold alignment, which preserves the hyperspherical geometry of the teacher embedding space. Self-supervised InfoNCE consistency and soft $\mathrm{SE}(3)$ invariance further encourage viewpoint-robust descriptors. In Stage 2, the distilled representation is adapted to registration through correspondence learning, density-aware point-dropout augmentation, and end-to-end pose optimization. With a single checkpoint, CVSD-Reg generalizes to both single-sensor and zero-shot cross-sensor scenarios without sensor-specific adaptation and remains entirely camera-free at inference. On KITTI, nuScenes, and HeLiPR, CVSD-Reg achieves strict success rate (SR@0.5\,m/$1^\circ$) of 97.7$\%$, 99.0$\%$, and 99.3$\%$, respectively, including 97.3$\%$ on sparse 16-beam Velodyne scans. It outperforms state-of-the-art geometric registration methods by up to 44.0 percentage points without requiring camera inputs or post-hoc ICP refinement.

Retrieval-augmented generation (RAG) improves language-model responses by grounding generation in external passages, which comes with overhead: retrieved context lengthens the prompt, increasing prefill work, KV-cache footprint, memory traffic, latency, and energy. Context compression offers a natural remedy by pruning retrieved text before generation. However, state-of-the-art context-compression methods are typically used with a fixed compression budget, or with the rate selected offline and then applied at inference time. This static view ignores both workload variation and the live state of the edge device. On an edge SoC, compression is not free: the compressor itself runs on the same SoC and consumes latency and energy that can offset any generation savings. This paper proposes a vision for telemetry-informed adaptive compression in edge RAG, grounded in experimental evidence. We characterize the compression tradeoff on the NVIDIA Jetson AGX Thor using Llama and Qwen generators, Natural Questions and HotpotQA datasets, and LLMLingua-2 compression. Our measurements show that generation dominates the RAG budget for larger models, reaching roughly 90% of per-query latency and 91% of GPU energy for 7B-8B generators. Exploring the impact of the compression rate reveals an adaptive operating region: mild compression can miss energy opportunities, and overly aggressive compression can hurt inference quality. Intermediate compression can reduce GPU energy by up to 53.2%, and SoC energy by up to 48.2%, with negligible quality loss. We argue for runtime policies that dynamically manage compression, guided by workload features and edge telemetry.

Gradient inversion attacks recover private training text from gradients shared in federated learning, posing a serious threat to collaborative model training. Through our analysis of transformer gradient structure, we identify three channels through which private token information leaks: the attention output projection gradient exposes a low-rank subspace that encodes input embeddings (Channel 1), the embedding gradient's row-norm sparsity directly reveals which tokens are present (Channel 2), and the MLP expansion gradient carries a recoverable subspace signal analogous to Channel 1 (Channel 3). State-of-the-art attacks exploit these channels analytically to achieve near-exact token recovery in seconds. Existing defences address at most one channel and either degrade model utility or leave the remaining structural signals intact. We introduce AEGIS (Attention-Embedding Gradient Isolation Shield), a lightweight defence that closes all three analytical channels with three backward-path operations requiring no architectural changes: freezing attention projection parameters eliminates Channel 1 by construction, calibrated noise injection into the embedding gradient destroys Channel 2's token-presence signal, and analogous per-block noise injection into the MLP expansion gradient masks Channel 3. The same masked gradient drives both the local optimiser step and the server export, so no clean signal is retained on either side. Evaluated across 11 models and six datasets, AEGIS reduces token recovery rates to near zero against a range of gradient inversion attacks, both analytical and optimisation-based, while preserving or improving model utility. We provide formal guarantees for Channels 1 and 2 and validate the full defence empirically against adaptive adversaries with complete knowledge of the mechanism.

Many real-world AI systems represent entities, behaviors, and structured information using discrete machine-native symbols rather than natural language. While these representations are compact and preserve task-relevant structure, they lie outside the linguistic token space of pretrained large language models (LLMs), creating a fundamental divide between language modeling and structured prediction. We introduce UniLang, a unified generative framework that bridges this divide by extending pretrained LLMs to treat machine-native symbols as first-class generative units alongside natural-language tokens. UniLang expands the LLM's vocabulary and embedding space with grounded machine-native representations, enabling textual and symbolic tokens to be jointly modeled and generated under a single autoregressive objective. This unified interface allows pretrained LLMs to directly operate on machine-native representations without requiring them to be verbalized as natural language or relying on task-specific architectures. We evaluate UniLang on two structurally distinct tasks, sequential recommendation and legal precedent prediction, spanning different domains and types of structured prediction. Across both tasks, UniLang consistently outperforms strong baselines, demonstrating a path toward extending pretrained LLMs beyond language and using them as a common generative modeling backbone for heterogeneous machine-native representations.

AI clones that imitate a specific person typically reproduce what the person says and how they sound, but not how they listen. We investigate whether adding multimodal listening behaviors gives such a clone more presence and authenticity. We integrated verbal backchannels and head nodding, driven by real-time prediction models, into an AI clone equipped with voice cloning and LLM-based responses. In a within-subjects study (N=35), adding these behaviors significantly improved the perceived attentiveness of the avatar, the sense of talking with the real person, and the feeling of co-presence. These results indicate that AI clone fidelity should extend beyond voice and response content to include interactive listening behavior.