ICLR 2024 accepted papers on Time series
52 of the 2,259 papers accepted at ICLR 2024 (2.3%) are on Time series, compared with 1.5% at ICLR 2023. They fall into 4 topics.
See ICLR 2024 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Causal and generative time series · 19 papers
- Diffusion-TS: Interpretable Diffusion for General Time Series Generation (12 citations)
- EasyTPP: Towards Open Benchmarking Temporal Point Processes (6 citations)
- Stable Neural Stochastic Differential Equations in Analyzing Irregular Time Series Data (6 citations)
- Copula Conformal prediction for multi-step time series prediction (4 citations)
- TACTiS-2: Better, Faster, Simpler Attentional Copulas for Multivariate Time Series (2 citations)
- Contrastive Difference Predictive Coding (1 citation)
- On the Generalization and Approximation Capacities of Neural Controlled Differential Equations (1 citation)
- Parallelizing non-linear sequential models over the sequence length (1 citation)
- RobustTSF: Towards Theory and Design of Robust Time Series Forecasting with Anomalies (1 citation)
- Efficient Inverse Multiagent Learning
- Let's do the time-warp-attend: Learning topological invariants of dynamical systems
- MG-TSD: Multi-Granularity Time Series Diffusion Models with Guided Learning Process
and 7 more in the interactive map.
Foundation models for time series · 18 papers
- TEMPO: Prompt-based Generative Pre-trained Transformer for Time Series Forecasting (36 citations)
- TEST: Text Prototype Aligned Embedding to Activate LLM's Ability for Time Series (20 citations)
- Learning to Embed Time Series Patches Independently (6 citations)
- Soft Contrastive Learning for Time Series (4 citations)
- Inherently Interpretable Time Series Classification via Multiple Instance Learning (2 citations)
- T-Rep: Representation Learning for Time Series using Time-Embeddings (2 citations)
- Parametric Augmentation for Time Series Contrastive Learning (1 citation)
- STanHop: Sparse Tandem Hopfield Model for Memory-Enhanced Time Series Prediction (1 citation)
- Towards Enhancing Time Series Contrastive Learning: A Dynamic Bad Pair Mining Approach (1 citation)
- AutoCast++: Enhancing World Event Prediction with Zero-shot Ranking-based Context Retrieval
- Explaining Time Series via Contrastive and Locally Sparse Perturbations
- Feature-aligned N-BEATS with Sinkhorn divergence
and 6 more in the interactive map.
Long-term forecasting · 14 papers
- iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (385 citations)
- TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting (93 citations)
- Pathformer: Multi-scale Transformers with Adaptive Pathways for Time Series Forecasting (37 citations)
- CARD: Channel Aligned Robust Blend Transformer for Time Series Forecasting (4 citations)
- Rethinking Channel Dependence for Multivariate Time Series Forecasting: Learning from Leading Indicators (4 citations)
- Self-Supervised Contrastive Learning for Long-term Forecasting (4 citations)
- ARM: Refining Multivariate Forecasting with Adaptive Temporal-Contextual Learning
- Biased Temporal Convolution Graph Network for Time Series Forecasting with Missing Values
- DAM: Towards a Foundation Model for Forecasting
- GAFormer: Enhancing Timeseries Transformers Through Group-Aware Embeddings
- Interpretable Sparse System Identification: Beyond Recent Deep Learning Techniques on Time-Series Prediction
- ModernTCN: A Modern Pure Convolution Structure for General Time Series Analysis
and 2 more in the interactive map.
Anomaly detection · 1 paper
- FITS: Modeling Time Series with $10k$ Parameters (17 citations)
