Unified Work Embeddings: Contrastive Learning of a Bidirectional Multi-task Ranker

2025-11-11

AI Summary

Existing general-purpose embedding models exhibit inconsistent performance and lack standardized evaluation under realistic workplace challenges—including long-tailed label distributions, extreme multi-label scenarios, and data scarcity. Method: We introduce WorkBench, the first comprehensive evaluation suite covering six distinct workplace tasks. We propose the Unified Workplace Embedding (UWE) model, a bidirectional multi-task ranking framework integrating graph-structure enhancement, many-to-many InfoNCE contrastive learning, a dual-encoder architecture, and token-level embeddings—augmented by synthetic data generation and bipartite graph modeling. UWE enables task-agnostic zero-shot ranking and target-space embedding caching for low-latency inference. Results: On WorkBench, UWE significantly outperforms general-purpose embedding models, achieving substantial gains in macro-average Mean Average Precision (MAP) and Recall-Precision@10 (RP@10). These results validate effective cross-task knowledge transfer and practical deployability in real-world workplace applications.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Humans and AI: Crowd Sourcing and Human Computation

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Workforce transformation across diverse industries has driven an increased demand for specialized natural language processing capabilities. Nevertheless, tasks derived from work-related contexts inherently reflect real-world complexities, characterized by long-tailed distributions, extreme multi-label target spaces, and scarce data availability. The rise of generalist embedding models prompts the question of their performance in the work domain, especially as progress in the field has focused mainly on individual tasks. To this end, we introduce WorkBench, the first unified evaluation suite spanning six work-related tasks formulated explicitly as ranking problems, establishing a common ground for multi-task progress. Based on this benchmark, we find significant positive cross-task transfer, and use this insight to compose task-specific bipartite graphs from real-world data, synthetically enriched through grounding. This leads to Unified Work Embeddings (UWE), a task-agnostic bi-encoder that exploits our training-data structure with a many-to-many InfoNCE objective, and leverages token-level embeddings with task-agnostic soft late interaction. UWE demonstrates zero-shot ranking performance on unseen target spaces in the work domain, enables low-latency inference by caching the task target space embeddings, and shows significant gains in macro-averaged MAP and RP@10 over generalist embedding models.

Problem

• Evaluates embedding models on work-related tasks with ranking problems
• Addresses data scarcity and complex distributions in workforce NLP tasks
• Develops unified embeddings for zero-shot ranking in specialized work domains

Innovation

• Contrastive learning with many-to-many InfoNCE objective
• Token-level embeddings with soft late interaction
• Bipartite graph data structure synthetically enriched