Interest Loss

Interest loss, in the context of machine learning, refers to the function used to quantify the difference between a model's predictions and the true values, guiding model optimization. Current research focuses on developing novel loss functions tailored to specific tasks and datasets, often incorporating geometric properties of embedding spaces or addressing issues like class imbalance and data scarcity. These advancements aim to improve model accuracy, efficiency, and robustness, impacting various applications from image recognition and speech enhancement to reinforcement learning and recommendation systems. The development of improved loss functions is crucial for pushing the boundaries of machine learning performance and addressing limitations in existing models.

Papers

November 5, 2024

On the loss of context-awareness in general instruction fine-tuning
Yihan Wang, Andrew Bai, Nanyun Peng, Cho-Jui Hsieh
Downstream Task Interest Loss Instruction Fine Tuning Context Awareness Instruction Response Pair

October 30, 2024

DASH: Warm-Starting Neural Network Training in Stationary Settings without Loss of Plasticity
Baekrok Shin, Junsoo Oh, Hanseul Cho, Chulhee Yun
Neural Network Interest Loss Plasticity Rule Warm Start Stationary Distribution Plasticity Loss Practical Deep

October 27, 2024

Prototypical Extreme Multi-label Classification with a Dynamic Margin Loss
Kunal Dahiya, Diego Ortego, David Jiménez
Soft Label Interest Loss Margin Maximization Label Space Extreme Multi Label Label Representation

October 7, 2024

SimO Loss: Anchor-Free Contrastive Loss for Fine-Grained Supervised Contrastive Learning
Taha Bouhsine, Imad El Aaroussi, Atik Faysal, Wang Huaxia
Jina Embeddings Contrastive Loss Interest Loss Post Hoc Orthogonalization Structural Embeddings Fine Grained Contrastive Learning

October 4, 2024

HyperCMR: Enhanced Multi-Contrast CMR Reconstruction with Eagle Loss
Ruru Xu, Caner Özer, Ilkay Oksuz
Cardiac Magnetic Resonance Interest Loss

August 8, 2024

Exploiting Consistency-Preserving Loss and Perceptual Contrast Stretching to Boost SSL-based Speech Enhancement
Muhammad Salman Khan, Moreno La Quatra, Kuo-Hsuan Hung, Szu-Wei Fu, Sabato Marco Siniscalchi, Yu Tsao
Speech Enhancement Self Supervised Representation Learning Interest Loss Mask Transformer Preserving Loss

July 9, 2024

It's Our Loss: No Privacy Amplification for Hidden State DP-SGD With Non-Convex Loss
Meenatchi Sundaram Muthu Selva Annamalai
PRIvacy Leakage Interest Loss Convex Loss Private Stochastic Gradient Descent Privacy Amplification

June 25, 2024

Beyond Silence: Bias Analysis through Loss and Asymmetric Approach in Audio Anti-Spoofing
Hye-jin Shim, Md Sahidullah, Jee-weon Jung, Shinji Watanabe, Tomi Kinnunen
Training Dynamic Bias Evaluation Interest Loss Spoofed Speech Audio Anti Spoofing Invisible Attack Asymmetric Learning

June 17, 2024

Do Parameters Reveal More than Loss for Membership Inference?
Anshuman Suri, Xiao Zhang, David Evans
Membership Inference Attack Many Parameter White Box Membership Inference Interest Loss Membership Privacy Hessian Inverse

June 9, 2024

Region of Interest Loss for Anonymizing Learned Image Compression
Christoph Liebender, Ranulfo Bezerra, Kazunori Ohno, Satoshi Tadokoro
Foundation Model Supervised Autoencoder Image Compression Region Specific Interest Loss Consistent Anonymization Effect

June 4, 2024

An Independence-promoting Loss for Music Generation with Language Models
Jean-Marie Lemercier, Simon Rouard, Jade Copet, Yossi Adi, Alexandre Défossez
Language Model Music Generation Auto Encoder Interest Loss

May 25, 2024

Improving 3D Occupancy Prediction through Class-balancing Loss and Multi-scale Representation
Huizhou Chen, Jiangyi Wang, Yuxin Li, Na Zhao, Jun Cheng, Xulei Yang
Semantic Segmentation 3D Occupancy Interest Loss Multi Scale Representation 3D Occupancy Prediction 3D Object Recognition Occupancy Mapping

March 8, 2024

Switching the Loss Reduces the Cost in Batch Reinforcement Learning
Alex Ayoub, Kaiwen Wang, Vincent Liu, Samuel Robertson, James McInerney, Dawen Liang, Nathan Kallus, Csaba Szepesvári
Hidden CoST Interest Loss Near Optimal Policy Batch Reinforcement Learning Fitted Q Iteration Small Loss

February 29, 2024

A Model-Based Approach for Improving Reinforcement Learning Efficiency Leveraging Expert Observations
Erhan Can Ozcan, Vittorio Giammarino, James Queeney, Ioannis Ch. Paschalidis
Deep Reinforcement Learning Behavior Cloning Model Based Interest Loss Expert Level Performance Expert Observation

February 24, 2024

Large Stepsize Gradient Descent for Logistic Loss: Non-Monotonicity of the Loss Improves Optimization Efficiency
Jingfeng Wu, Peter L. Bartlett, Matus Telgarsky, Bin Yu
Gradient Descent Neural Tangent Kernel Efficient Optimization Interest Loss Logistics System Planning Loss Nonlinear Prediction Partial Monotonicity

December 21, 2023

On the Convergence of Loss and Uncertainty-based Active Learning Algorithms
Daniel Haimovich, Dima Karamshuk, Fridolin Linder, Niek Tax, Milan Vojnovic
Loss Function Stochastic Gradient Descent Early Stage Convergence Interest Loss Uncertainty Based Active Learning Smooth Loss Function

December 19, 2023

Loss it right: Euclidean and Riemannian Metrics in Learning-based Visual Odometry
Olaya Álvarez-Tuñón, Yury Brodskiy, Erdal Kayacan
Visual Odometry Interest Loss Riemannian Metric Pose Representation Learning Based Visual Odometry

November 30, 2023

Directions of Curvature as an Explanation for Loss of Plasticity
Alex Lewandowski, Haruto Tanaka, Dale Schuurmans, Marlos C. Machado
Neural Network Supervised ImageNet Line by Line Explanation Interpretable Direction Interest Loss Plasticity Rule Gaussian Curvature MNIST Canadian Institute for Advanced Continual Learning Task

November 28, 2023

iMagLS: Interaural Level Difference with Magnitude Least-Squares Loss for Optimized First-Order Head-Related Transfer Function
Or Berebi, Zamir Ben-Hur, David Lou Alon, Boaz Rafaely
Interest Loss Spatial Audio Head Related Transfer Function Earthquake Magnitude Order Ambisonics Binaural Rendering Inter Aural

November 22, 2023

A Joint Gradient and Loss Based Clustered Federated Learning Design
Licheng Lin, Mingzhe Chen, Zhaohui Yang, Yusen Wu, Yuchen Liu
Interest Loss Joint Optimization Local Model Clustered Federated Learning FL Algorithm Efficient Clustering

Interest Loss

Papers

On the loss of context-awareness in general instruction fine-tuning

DASH: Warm-Starting Neural Network Training in Stationary Settings without Loss of Plasticity

Prototypical Extreme Multi-label Classification with a Dynamic Margin Loss

SimO Loss: Anchor-Free Contrastive Loss for Fine-Grained Supervised Contrastive Learning

HyperCMR: Enhanced Multi-Contrast CMR Reconstruction with Eagle Loss

Exploiting Consistency-Preserving Loss and Perceptual Contrast Stretching to Boost SSL-based Speech Enhancement

It's Our Loss: No Privacy Amplification for Hidden State DP-SGD With Non-Convex Loss

Beyond Silence: Bias Analysis through Loss and Asymmetric Approach in Audio Anti-Spoofing

Do Parameters Reveal More than Loss for Membership Inference?

Region of Interest Loss for Anonymizing Learned Image Compression

An Independence-promoting Loss for Music Generation with Language Models

Improving 3D Occupancy Prediction through Class-balancing Loss and Multi-scale Representation

Switching the Loss Reduces the Cost in Batch Reinforcement Learning

A Model-Based Approach for Improving Reinforcement Learning Efficiency Leveraging Expert Observations

Large Stepsize Gradient Descent for Logistic Loss: Non-Monotonicity of the Loss Improves Optimization Efficiency

On the Convergence of Loss and Uncertainty-based Active Learning Algorithms

Loss it right: Euclidean and Riemannian Metrics in Learning-based Visual Odometry

Directions of Curvature as an Explanation for Loss of Plasticity

iMagLS: Interaural Level Difference with Magnitude Least-Squares Loss for Optimized First-Order Head-Related Transfer Function

A Joint Gradient and Loss Based Clustered Federated Learning Design