archive
Every paper Pith has read. Search by title, abstract, or pith.
74322 papers indexed · page 42
-
Hidden states at paragraph boundaries tune verifier strictness
The Hidden Signal of Verifier Strictness: Controlling and Improving Step-Wise Verification via Selective Latent Steering
-
Testbed embeds detectable hacks for automatic reward-gaming checks
Hack-Verifiable Environments: Towards Evaluating Reward Hacking at Scale
-
Constraint engine turns AI drawings into verifiable geometry reasoning
Draw2Think: Harnessing Geometry Reasoning through Constraint Engine Interaction
-
Text modeling of EV battery signals enables LLM fault diagnosis
VBFDD-Agent for Electric Vehicle Battery Fault Detection and Diagnosis: Descriptive Text Modeling of Battery Digital Signals
-
Mid-IR imaging system delivers 25 mm field with edge enhancement
Wide-field mid-infrared edge-enhanced upconversion imaging
-
RL scores full distributions to fix LLM regression
Distribution-Aware Reward: Reinforcement Learning over Predictive Distributions for LLM Regression
-
Maximizing naive bounds recovers misspecified Cramér-Rao bound
Revisiting the Misspecified Cram\'er-Rao Bound
-
Scale-decoupled alignment improves remote sensing incremental detection
STAR-IOD: Scale-decoupled Topology Alignment with Pseudo-label Refinement for Remote Sensing Incremental Object Detection
-
Language priors fix long-tail bias in 3D point cloud clustering
Resolving Long-Tail Ambiguity in Unsupervised 3D Point Cloud Segmentation with Language Priors
-
Addition theorems yield exact elastic potential matrices
Addition Theorems for Real Vector Spherical Harmonics and Explicit Matrix Representations of the Quasi-Periodic Elastic Single Layer Potential
-
Open-source iris algorithms pass first official IREX evaluation
Lowering the Barrier to IREX Participation: Open-Source Algorithms, Toolkit, and Benchmarking for Iris Recognition
-
Monitor reduces LLM agent covert channels to zero capacity
An Application-Layer Multi-Modal Covert-Channel Reference Monitor for LLM Agent Egress
-
Method generates editable 3D surfaces from hand sketches
Sketch2MinSurf: Vision-Language Guided Generation of Editable Minimal Surfaces from Hand-Drawn Sketches
-
Attention reweighting suppresses spurious features before CNN pooling
Deep Attention Reweighting: Post-Hoc Attention-Based Feature Aggregation in CNNs for Disentangling Core and Spurious Features under Spurious Correlations
-
Designer ratings dataset lifts AI graphic scorer to 0.611 agreement
TASTE: A Designer-Annotated Multi-Dimensional Preference Dataset for AI-Generated Graphic Design
-
Aligning task vectors to in-context next-token distributions lifts accuracy 9.2%
Distributional Alignment as a Criterion for Designing Task Vectors in In-Context Learning
-
Framework synthesizes realistic conversational retrieval benchmarks
MTR-Suite: A Framework for Evaluating and Synthesizing Conversational Retrieval Benchmarks
-
Early high-frequency injection reduces OOD score overlap
Early High-Frequency Injection for Geometry-Sensitive OOD Detection
-
Virtual outliers reshape geometry to handle noisy labels
GAMR: Geometric-Aware Manifold Regularization with Virtual Outlier Synthesis for Learning with Noisy Labels
-
Conformal tests bound false discoveries for every possible threshold
Everywhere Valid Bounds on False Discovery Proportions in Conformal Inference
-
Decoupling reliabilities lifts noisy-label accuracy
Holistic Reliability Propagation: Decoupling Annotation and Prediction for Robust Noisy-Label
-
Dual memory layers give LLMs unbounded conversation context
CALMem : Application-Layer Dual Memory for Conversational AI
-
Android crowds run large DNNs at 43 MB RAM per phone
Memory-Efficient Partitioned DNN Inference on Resource-Constrained Android Crowds
-
Group statistics adapt clipping and temperature to lift LLM math scores
AGPO: Adaptive Group Policy Optimization with Dual Statistical Feedback
-
Functor l preserves tilting pairs in cleft extensions
Tilting pairs and Wakamatsu tilting pairs of subcategories over cleft extensions
-
GMM calibration lets recommenders use all noisy feedback
Robust Recommendation from Noisy Implicit Feedback: A GMM-Weighted Bayes-label Transition Matrix Framework
-
Non-elliptic terms give o(X) errors in summed GL2 trace formula
Beyond endoscopy for $\mathsf{GL}_2$ over $\mathbb{Q}$ with ramification 4: contribution of non-elliptic parts
-
Explicit Gâteaux formula yields mean-field policy gradient
Policy Gradient for Continuous-Time Mean-Field Control
-
ReRAM macro reaches 419 TOPS/W for edge neural inference
E-ReCON: An Energy- and Resource-Efficient Precision-Configurable Sparse nvCIM Macro for Conventional and Spiking Neural Edge Inference
-
Decision path flips raise random forest accuracy
Decision-Path Patterns as Tree Reliability Signals: Path-based Adaptive Weighting for Random Forest Classification
-
Decision-path flips yield unbiased per-sample weights for random forests
Decision-Path Patterns as Tree Reliability Signals: Path-based Adaptive Weighting for Random Forest Classification
-
Scattering-blowup split holds at energy threshold for 4d NLS
Threshold dynamics for the 4$d$ mass-energy double critical NLS
-
Multiple formats complicate astronomy visualization tools
Data Formats and Visualisation BoF
-
SAVER selectively activates vision to boost F1 and cut latency in multimodal IE
SAVER: Selective As-Needed Vision Evidence for Multimodal Information Extraction
-
Categorical error rates beat WER for Indic speech recognition
SCRIBE: Diagnostic Evaluation and Rich Transcription Models for Indic ASR
-
New algorithm respects constraint priorities even when infeasible
Augmented Lagrangian methods for convex optimization with priority constraints via an infeasibility control framework
-
New method checks observational CATE predictions against trial results
Assessing Estimate of CATE from Observational Data via an RCT Study
-
35% of non-periodic packings show selective order
What Lies Between Crystal and Randomly Packed Structures? A General Characterization of Non-Periodic Order
-
DAR cuts DiT training iterations by 8.75x while improving FID by 2.11
Rethinking Cross-Layer Information Routing in Diffusion Transformers
-
GL(3) Fourier sums scaled by x^{-1/3} have a limiting distribution
Limiting Distribution and Rate of Convergence for GL(3) Fourier Coefficients
-
WebGPU backend cuts LLM memory use by 29-33% in browsers
Llamas on the Web: Memory-Efficient, Performance-Portable, and Multi-Precision LLM Inference with WebGPU
-
Forbidden s-point pattern forces sub-maximal incidences
Extremal structure in dense arrangements of $k$-intersecting curves
-
Heartbeat protocol revokes AI swarm credentials within fixed window
Heartbeat-Bound Hierarchical Credentials: Cryptographic Revocation for AI Agent Swarms
-
io-HEOM captures non-Markovian waveguide QED from two sources
Strongly-coupled non-Markovian waveguide QED with input-output HEOM
-
LLMs endorse 32% of their own behavior-changing code rewrites
Articulate but Wrong: Self-Review Failures in LLM-Based Code Modernization
-
Randomized Chirikov map mixes exponentially almost surely
Quantitative exponential mixing for the randomized Chirikov standard map
-
AI simulator trains clinicians on disclosing medical errors
CandorMD: An AI-Assisted Audio Simulation and Feedback System for Training Clinicians for Medical Error Disclosure
-
Gaussian limit holds for maxima under weaker correlation decay
The maximum of a strongly correlated Gaussian process
-
Substrate dipoles set 2D nonlinear conductivity to 1 μm/ΩV
Giant nonlinear conductivity in 2D electron gas from substrate-induced dipolar scattering
-
Quasar ISM line ratios require heating beyond standard PDR models
CO(7-6) and [C I](2-1) survey in z > 6 quasars