New Tracks at EMNLP 2025 and Their Relationship to ARR Tracks

As we prepare for EMNLP 2025, many of you may have noticed that we are introducing several new tracks and making adjustments to our area list. You might also notice that there are some differences between the EMNLP tracks and those currently reflected in the ACL Rolling Review (ARR) system. Note that the EMNLP tracks will apply only at the commitment stage, i.e. after the ARR reviewing cycle is finished. At initial submission time you will need to choose among the ARR tracks. To help with this transition, the ARR area keywords have been updated to better align with the new EMNLP tracks. You can search for your preferred sub-areas and their associated ARR track here: https://aclrollingreview.org/areas. We wanted to take a moment to explain the motivations behind these decisions and how they fit into the broader picture of our field’s evolution.

Why We Are Adding New Tracks

The field of NLP is rapidly evolving. Areas that were once niche have grown into major subfields, and entirely new topics have emerged. To encourage cutting-edge research, EMNLP 2025 will:

  • Create space and foster community for fast-growing and emerging research areas that deserve focused attention.

  • Experiment with track structure in a way that informs future ARR reforms.
    EMNLP, as a one-off event, gives us the flexibility to pilot changes that can later help ARR evolve thoughtfully and responsibly.

While ARR needs to maintain long-term stability, commitment stages of individual conferences can be more agile – and the lessons from this conference will provide valuable feedback and inform more permanent changes to the ARR tracks going forward.

What’s Being Added and The Relationship with ARR Existing Tracks

Here’s a summary of the new or adjusted tracks, organized by the reasoning behind them:

1. Recognizing Emerging Research Areas

We are creating or highlighting tracks to reflect research areas that have recently expanded rapidly:

  • Safety and Alignment in LLMs
  • AI/LLM Agents
  • Code Models
  • Neurosymbolic Approaches to NLP
  • Mathematical, Symbolic, and Logical Reasoning
  • Generalizability and Transfer
  • LLM Efficiency

2. Expanding/Clarifying Existing Areas

To clarify reviewer expertise and topic boundaries, we also propose the following changes:

  • Information Extraction and Retrieval
    Merged from the tracks of information extraction and Information Retrieval and Text Mining tracks

  • Interpretability, Model Editing, and Explainability:
    Expanded from the Interpretability and Analysis track

  • Hierarchical Structure Prediction, Syntax, and Parsing:
    Added hierarchical structure prediction as a keyword under the Syntax and Parsing track to better reflect current research directions.

3. Mapping to ARR Tracks

The following EMNLP 2025 tracks will be organized as sub-areas under the ARR track Language Modeling:

  • Safety and Alignment in LLMs
  • AI/LLM Agents
  • Code Models
  • Neurosymbolic Approaches to NLP

The following EMNLP 2025 tracks will be organized as sub-areas under the ARR track Machine Learning for NLP:

  • Mathematical, Symbolic, and Logical Reasoning
  • Generalizability and Transfer

The EMNLP 2025 track of LLM Efficiency will be organized as sub-areas under the ARR track Efficient/Low-Resource Methods for NLP.

The EMNLP 2025 track of Information Extraction and Retrieval essentially merged the ARR tracks of Information Extraction and Information Retrieval and Text Mining.

Closing Remarks

The track updates at EMNLP 2025 are more than just adjustments for this year — they are an opportunity to proactively shape the future of our community.

By experimenting with new tracks and structures now, we hope to gain valuable insights into submission patterns, reviewing quality, and community needs — insights that can inform the evolution of the ACL Rolling Review for the better.

We look forward to your submissions and to working together to build the next generation of conferences!

Updated: