Multi Task Inference

Multi-task inference aims to perform multiple prediction tasks simultaneously using a single model, improving efficiency and potentially accuracy compared to separate models for each task. Current research focuses on developing efficient algorithms and architectures, such as parameter-efficient fine-tuning methods and memory-efficient pruning techniques, to handle diverse tasks within resource-constrained environments, including edge devices and UAVs. This approach is particularly relevant for large language models and computer vision, offering benefits in speed, energy consumption, and generalization across various applications like autonomous driving and natural language processing. The improved efficiency and potential performance gains are driving significant interest in this area.

Papers

October 3, 2024

FastAdaSP: Multitask-Adapted Efficient Inference for Large Speech Language Model
Yichen Lu, Jiaqi Song, Chao-Han Huck Yang, Shinji Watanabe
Language Model Question Answering Efficient Inference Language Model Inference Multi Task Inference Multitask Speech

October 1, 2024

Embedding-based statistical inference on generative models
Hayden Helm, Aranyak Acharyya, Brandon Duderstadt, Youngser Park, Carey E. Priebe
Generative Model Generative AI Model Relevant Covariates Statistical Inference Multi Task Inference

September 16, 2024

Human Insights Driven Latent Space for Different Driving Perspectives: A Unified Encoder for Efficient Multi-Task Inference
Huy-Dung Nguyen, Anass Bairouk, Mirjana Maras, Wei Xiao, Tsun-Hsuan Wang, Patrick Chareyre, Ramin Hasani, Marc Blanchon, Daniela Rus
Autonomous Driving Latent Space Steering Technique Driving Scenario Multi Task Inference Steering Angle Prediction

May 15, 2024

Compressive Feature Selection for Remote Visual Multi-Task Inference
Saeed Ranjbar Alvar, Ivan V. Bajić
Mutual Information Feature Compression Feature Sparsity Multi Task Inference Remote Inference

February 18, 2024

Multi-Task Inference: Can Large Language Models Follow Multiple Instructions at Once?
Guijin Son, Sangwon Baek, Sangdae Nam, Ilgyun Jeong, Seungone Kim
Large Language Model Inference Time Inference Framework Diverse Instruction Multi Task Inference

December 31, 2023

Energy-Efficient Power Control for Multiple-Task Split Inference in UAVs: A Tiny Learning-Based Approach
Chenxi Zhao, Min Sheng, Junyu Liu, Tianshu Chu, Jiandong Li
Unmanned Aerial Vehicle Tiny Machine Learning Multi Task Inference

October 30, 2023

Res-Tuning: A Flexible and Efficient Tuning Paradigm via Unbinding Tuner from Backbone
Zeyinzi Jiang, Chaojie Mao, Ziyuan Huang, Ao Ma, Yiliang Lv, Yujun Shen, Deli Zhao, Jingren Zhou
Parameter Efficient Tuning Hyper Tune Automatic Tuning Numerous Cutting Edge Backbone Multi Task Inference Tuning Method

May 18, 2023

Ahead-of-Time P-Tuning
Daniil Gavrilov, Nikita Balagansky
Pre Trained Language Model Efficient Fine Tuning Early Prediction Multi Task Inference P Tuning

March 25, 2023

Energy-efficient Task Adaptation for NLP Edge Inference Leveraging Heterogeneous Memory Architectures
Zirui Fu, Aleksandre Avaliani, Marco Donato
Edge Device Task Adaptation Transformer Based Deep Edge Inference AI Accelerator Multi Task Inference Heterogeneous Memory

December 21, 2022

SPT: Semi-Parametric Prompt Tuning for Multitask Prompted Learning
M Saiful Bari, Aston Zhang, Shuai Zheng, Xingjian Shi, Yi Zhu, Shafiq Joty, Mu Li
Large Language Model Prompt Tuning Multitask Learning Multi Task Inference

May 2, 2022

Paragraph-based Transformer Pre-training for Multi-Sentence Inference
Luca Di Liello, Siddhant Garg, Luca Soldaini, Alessandro Moschitti
Fact Checking Pre Trained Transformer Inference Task Multi Task Inference Sentence Inference

April 11, 2022

MIME: Adapting a Single Neural Network for Multi-task Inference with Memory-efficient Dynamic Pruning
Abhiroop Bhattacharjee, Yeshwanth Venkatesha, Abhishek Moitra, Priyadarshini Panda
Multi Task Learning Single Neural Network Energy Efficient Inference Random Access Memory Multi Task Inference

February 8, 2022

Generative multitask learning mitigates target-causing confounding
Taro Makino, Krzysztof J. Geras, Kyunghyun Cho
Multi Task Learning Multi Task Multitask Learning Faulty Negative Mitigation Causal Representation Multi Task Inference

Multi Task Inference

Papers

FastAdaSP: Multitask-Adapted Efficient Inference for Large Speech Language Model

Embedding-based statistical inference on generative models

Human Insights Driven Latent Space for Different Driving Perspectives: A Unified Encoder for Efficient Multi-Task Inference

Compressive Feature Selection for Remote Visual Multi-Task Inference

Multi-Task Inference: Can Large Language Models Follow Multiple Instructions at Once?

Energy-Efficient Power Control for Multiple-Task Split Inference in UAVs: A Tiny Learning-Based Approach

Res-Tuning: A Flexible and Efficient Tuning Paradigm via Unbinding Tuner from Backbone

Ahead-of-Time P-Tuning

Energy-efficient Task Adaptation for NLP Edge Inference Leveraging Heterogeneous Memory Architectures

SPT: Semi-Parametric Prompt Tuning for Multitask Prompted Learning

Paragraph-based Transformer Pre-training for Multi-Sentence Inference

MIME: Adapting a Single Neural Network for Multi-task Inference with Memory-efficient Dynamic Pruning

Generative multitask learning mitigates target-causing confounding