CMU researchers are presenting 50 papers on the Thirtieth Convention on Empirical Strategies in Pure Language Processing (EMNLP 2025), held from November 4 – 9 in Suzhou, China. This contains 27 papers in the principle convention, 19 papers within the Findings observe, 2 system demonstrations papers, and a couple of business observe papers. This weblog submit offers aggregated details about EMNLP 2025 papers printed by CMU researchers.
Key areas addressed are visualized under (representing 30 of the 50 whole papers), illustrating the breadth of NLP and machine studying analysis being performed at CMU :
Observe: All info on this submit has been obtained by the ACL Anthology API and the EMNLP 2025 Presentation Data spreadsheet. Please contact CMU ML Weblog editors if you need any info added or modified.
Desk of Contents
Principal Convention Papers
Particular Theme: Interdisciplinary Recontextualization of NLP
Spontaneous Giving and Calculated Greed in Language Fashions
Artificial Socratic Debates: Analyzing Persona Results on Ethical Choice and Persuasion Dynamics
Multimodality and Language Grounding to Imaginative and prescient, Robotics and Past
Social Genome: Grounded Social Reasoning Skills of Multimodal Fashions
VisualWebInstruct: Scaling up Multimodal Instruction Knowledge by Net Search
Sources and Analysis
Persona-Augmented Benchmarking: Evaluating LLMs Throughout Various Writing Types
Human-AI Interplay/Cooperation
Estimating LLM Consistency: A Person Baseline vs Surrogate Metrics
Humanizing Machines: Rethinking LLM Anthropomorphism By way of a Multi-Degree Framework of Design
Interpretability, Mannequin Enhancing, Transparency, and Explainability
Calibrating LLMs for Textual content-to-SQL Parsing by Leveraging Sub-clause Frequencies
Mathematical, Symbolic, and Logical Reasoning in NLP
Rewarding the Unlikely: Lifting GRPO Past Distribution Sharpening
Agentic-R1: Distilled Twin-Technique Reasoning
Generalizability and Switch
SOCIAL SCAFFOLDS: A Generalization Framework for Social Understanding Duties
Looking for the Most Human-like Emergent Language
NLP Functions
PhoniTale: Phonologically Grounded Mnemonic Technology for Typologically Distant Language Pairs
Security and Alignment in LLMs
Anecdoctoring: Automated Crimson-Teaming Throughout Language and Place
Pure Language Technology
CIE: Controlling Language Mannequin Textual content Generations Utilizing Steady Indicators
Query Answering
Desk-R1: Inference-Time Scaling for Desk Reasoning Duties
Multilinguality and Language Variety
Grounding Multilingual Multimodal LLMs With Cultural Data
Computational Social Science, Cultural Analytics, and NLP for Social Good
Phrases Like Knives: Backstory-Customized Modeling and Detection of Violent Communication
AI/LLM Brokers
On the Tremendous-Grained Planning Skills of VLM Net Brokers
Code Fashions
An Empirical Examine on Robust-Weak Mannequin Collaboration for Repo-level Code Technology
Summarization
Summarizing Speech: A Complete Survey
Retrieval-Augmented Language Fashions
MoR: Higher Dealing with Various Queries with a Combination of Sparse, Dense, and Human Retrievers
Phonology, Morphology and Phrase Segmentation
Morpheme Induction for Emergent Language
Low-resource Strategies for NLP
Language Fashions Could be Effectively Steered through Minimal Embedding Layer Transformations
Findings Papers
Particular Theme: Interdisciplinary Recontextualization of NLP
FicSim: A Dataset for Multi-Faceted Semantic Similarity in Lengthy-Type Fiction
Sources and Analysis
SimBA: Simplifying Benchmark Evaluation Utilizing Efficiency Matrices Alone
mrCAD: Multimodal Communication to Refine Pc-aided Designs
Human-AI Interplay/Cooperation
Interpretability, Mannequin Enhancing, Transparency, and Explainability
Linear Steerability in Language Fashions: When It Emerges and How It Evolves
Predicting Language Fashions’ Success at Zero-Shot Probabilistic Prediction
Multilinguality and Language Variety
BenchMAX: A Complete Multilingual Analysis Suite for Giant Language Fashions
AI/LLM Brokers
FLAIRR-TS – Forecasting LLM-Brokers with Iterative Refinement and Retrieval for Time Collection
Code Fashions
VisCoder: Tremendous-Tuning LLMs for Executable Python Visualization Code Technology
Retrieval-Augmented Language Fashions
GAMIC: Graph-Aligned Molecular In-context Studying for Molecule Evaluation through LLMs
Speech Processing and Spoken Language Understanding
SVeritas: Benchmark for Sturdy Speaker Verification underneath Various Circumstances
CAARMA: Class Augmentation with Adversarial Mixup Regularization
Semantics: Lexical, Sentence-Degree Semantics, Textual Inference, and Different Areas
Bridging the Enhancing Hole in LLMs: FineEdit for Exact and Focused Textual content Modifications
Ethics, Bias, and Equity
Mitigate One, Skew One other? Tackling Intersectional Biases in Textual content-to-Picture Fashions
Dialogue and Interactive Techniques
LLM Effectivity
TreeBoN: Enhancing Inference-Time Alignment with Speculative Tree-Search and Finest-of-N Sampling
System Demonstrations
AgentDiagnose: An Open Toolkit for Diagnosing LLM Agent Trajectories
BioGraphia: A LLM-Assisted Organic Pathway Graph Annotation Platform
Trade Observe Papers
Leveraging LLMs to Streamline the Evaluate of Public Funding Functions
Semantic Settlement Permits Environment friendly Open-Ended LLM Cascades
