contrastive alignment

The papers

Most central to this idea first, not most recent.

• Big Techcs.CV

  Weak models can improve image-text alignment with statistical calibration

  Permute-then-Adapt uses permutation testing on weak model similarities to select auxiliary positives, improving image-text alignment with Type-I error control.

  The University of Melbourne, Southeast University · Nov 2026

• Big Techcs.CV

  AI avatars can copy face and body movements in real time

  A feed-forward transformer learns a compact motion embedding for real-time whole-human animation.

  NVIDIA, Seoul National University · Nov 2026

• Big Techcs.SD

  AI audio models better recognize individual sounds and whole soundscapes

  AudioFM achieves state-of-the-art audio representation learning by combining recording-level contrast with patch-wise masked latent discrimination.

  Google Research, Google DeepMind · Oct 2026

• Chinese Techcs.CV

  Many contrastive training strategies for multimodal embeddings add little value

  Most contrastive strategies yield marginal gains; combining hard-negative mining, false-negative mitigation, and hard-batch construction works best.

  Nanyang Technological University, Alibaba Group · Oct 2026

• Big Techcs.CV

  Balancing within-person and between-person variation improves clothes-changing re-identification

  MINER balances inter- and intra-person variance through stochastic hard mining to improve clothes-changing person re-identification.

  Reichman University, Microsoft · Oct 2026

• Independentcs.SD

  AI music dataset contains duplicate versions in over one-fifth of files

  At least 38,134 files in the Lakh MIDI dataset are duplicates, and filtered lists are provided.

  Nov 2026

• Chinese Techcs.CL

  AI annotators better reflect human labeling disagreements after learning from feedback

  Multi-agent contrastive framework ANT-OR boosts annotation agreement 4.5 points and matches human Fleiss kappa within 2 points.

  ANT Group, Harbin Institute of Technology (Shenzhen) · Oct 2026

• Big Techcs.CL

  Bi-encoders can support multimodal decisions across small and massive choice sets

  MetaEncoder turns a 30B multimodal decoder into a scalable bi-encoder that beats state-of-the-art multimodal embedders on decision tasks but remains weak on reasoning.

  Meta AI · Oct 2026

• Chinese Techcs.CV

  Human-rated clips improve rewards for synchronizing generated sound and video

  SyncReward is a human-aligned reward model that improves audio-visual synchronization in generated content.

  Renmin University of China, WeChat, Tencent Inc. · Oct 2026

• Research Labcs.CV

  Two same-image global views drive semantic structure in DINO models

  DINO-style semantic structure comes from aligning two geometrically distinct global crops of the same image; masking refines it, while local crops mainly improve classification.

  CISPA Helmholtz Center for Information Security, University of Freiburg · Oct 2026

• Big Techcs.CV

  MRI foundation model aligns anatomy and abnormalities across 3D volumes

  A 3D MRI vision-language model that aligns image-text representations at the anatomical level and separates healthy from abnormal anatomy outperforms existing MRI foundation models.

  GE Healthcare, Amazon · Oct 2026

• AI Startupcs.SD

  VoiceNet tests recognition of nuanced emotions and speaking styles

  VoiceNet benchmarks fine-grained voice understanding with 40 emotions and 57 talking-style attributes, and VoiceCLAP models outperform prior CLAP baselines on it.

  LAION e.V., Scalable Learning & Multi-Purpose AI (SLAMPAI) Lab · Oct 2026

• Top Universitycs.CV

  Multimodal AI models often ignore the mental images they generate during reasoning

  Latent visual states in multimodal LLMs often do not causally affect final answers.

  The University of Texas at Austin, UC Berkeley · Nov 2026

• Big Techcs.LG

  One compact model embeds text, images, audio and video together

  EmbeddingGemma 2 unifies text, images, audio, and video into a single embedding space, outperforming models two to twelve times larger.

  Google · Oct 2026

• Big Techcs.LG

  Pairing ECGs with multiple clinical report types yields a more universal ECG foundation model

  Training on multiple clinical note types improves ECG foundation model performance beyond using only ECG reports.

  Massachusetts Institute of Technology, Apple, Inc · Oct 2026

• Big Techcs.AI

  Safety instruction influence controlled via eigenvalue modulation in transformers

  Safety instructions in transformer LMs act as a multiplicative operator whose dominant eigenvalue controls instruction strength, enabling a contrastive loss to tune the attack/over-refusal tradeoff.

  Google Research, Cambridge University · Sept 2026

• Chinese Techcs.CV

  AI that identifies radiological findings first can diagnose unseen diseases

  Defining a disease as a fixed composition of learned radiological findings detects and localizes unseen and rare diseases better than predicting disease labels directly.

  University of California, Berkeley, Alibaba Group · Nov 2026

• Big Techcs.CL

  Learned prompts combine evolving LLM requirements without changing model weights

  Ready2Blend learns alignment requirements as composable prompts over a frozen backbone, matching post-training methods with less training time.

  Korea Advanced Institute of Science Technology, Samsung SDS · Oct 2026

• Big Techcs.LG

  Lightweight constitutional adapters improve jailbreak defenses without adversarial training

  Constitutional adapters, formed by subtracting control-trained from constitution-trained lightweight adapters, improve jailbreak defense and measured alignment without training on harmful requests.

  Anthropic Fellows Program, Independent · Sept 2026

• Top Universitycs.AI

  Anchor distributions define context-specific similarity in contrastive models

  Contrastive representations naturally contain a family of geometries that can be specialized by a distribution over semantic anchors.

  University of Chicago · Oct 2026

• Top Universitycs.LG

  Semantic alignment improves edge classification across unseen dynamic graph domains

  Spatio-Temporal Semantic Alignment outperforms all existing methods on edge classification for dynamic text-attributed graphs.

  Imperial College London · Sept 2026

• Chinese Techcs.CV

  Latent trajectory guidance improves video generation reasoning

  VR-JEPA uses latent trajectory prediction to guide video generation for improved visual reasoning.

  Sun Yat-sen University, Shenzhen Loop Area Institute · Oct 2026

• Top Universitycs.CV

  End-to-end model writes 3D scene graphs directly from point clouds or Gaussian splats

  GraphWrit3R directly converts point clouds or Gaussian Splats into JSON scene graphs and achieves state-of-the-art recall on 3DSSG without inference-time object annotations.

  INSAIT, Sofia University, Stanford University · Sept 2026

• Top Universitycs.CV

  Text queries retrieve SAE features for steering diffusion transformers

  D-Scope retrieves sparse autoencoder decoder directions for diffusion transformers by matching text queries to visual centroids of activating patches, enabling interpretable generation steering.

  Pivotal Research, Mohamed bin Zayed University of Artificial Intelligence · Oct 2026

• Chinese Tech

  Dynamic optimization loop aligns 3D generators with 2D diffusion priors to improve realism

  OREO enhances 3D generation fidelity by editing rendered views and distilling the gap back into the generator via latent contrastive learning.

  The Hong Kong Polytechnic University, Tencent ARC Lab · Oct 2026

• Chinese Techcs.CV

  Image search models can retain left-right information but ignore it

  A patch-grid readout restores left/right image search by recovering spatial relations that vision-language similarity either hides in pooled image summaries or discards during pooling.

  Huawei Technologies Canada · Nov 2026

• Big Techcs.CV

  Image forensics protocol sorts AI edits into retouching, tampering, and full synthesis

  A dual-channel protocol distinguishes retouched, tampered, and synthetic images beyond binary real/fake labels.

  Google, University of British Columbia · Nov 2026

• Academiccs.CV

  Video reasoning models improve by auditing their own use of visual clues

  A model can audit its own visual reliance and focus training on under-utilized visual tokens, improving video reasoning with far fewer resources.

  University of Texas at Austin · Nov 2026

• Big Techcs.RO

  Robots can reorient objects from pictures without tracking their pose

  A robot hand policy using only camera images and joint angles, with goals as images instead of object poses, reorients objects and transfers from simulation to real hands.

  ETH Zurich, NVIDIA · Nov 2026

• Big Techcs.CV

  Composed image retrieval gets a boost from factorized MLLM embeddings

  FactEmbed uses factorized latents to simulate visual modifications before retrieval, achieving state-of-the-art composed image retrieval.

  KAIST AI, Google DeepMind · Nov 2026

• Top Universitycs.CV

  Patch-token decoding does not reliably explain vision transformer decisions

  Token-decoding explainability is not an inherent property of vision transformers; it varies with training conditions.

  Technion – Israel Institute of Technology · Oct 2026

• Big Techcs.CV

  Unified visual encoder delivers state, language, and reconstruction for world models

  ALPS is a visual encoder for world models that jointly learns state information, language alignment, and pixel reconstruction from still images and text.

  University of California, Santa Cruz, University of California, Santa Barbara · Oct 2026

• Big Techcs.LG

  Minimal corrections teach hard tasks while preserving broader model skills

  Minimal, refreshed corrections of failed LLM attempts improve hard-task specialization while preserving prior capabilities better than on-policy training or full-answer replacement.

  AWS AI Labs, University of Massachusetts · Oct 2026

• Big Techcs.RO

  Action-aware embeddings retrieve better robot experiences for future actions

  Robota-Embed jointly encodes vision, language, and action for robot experience retrieval, outperforming action-free embeddings.

  NVIDIA, MIT · Oct 2026

• Big Techcs.CL

  Translation agents learn to retrieve evidence only when it improves results

  LLM translation improves by learning when and what to retrieve instead of always retrieving.

  The University of Texas at Austin, Amazon · Oct 2026

• Big Techcs.AI

  Text proxies can replace paired multimodal data for cross-modal reasoning

  Multimodal LLMs can learn cross-modal reasoning without paired non-text data by using proxy text as semantic bridges.

  University of California, Riverside, Amazon · Oct 2026

• Chinese Techcs.AI

  Safety training works better when models ignore superficial prompt cues

  DeShortcut-Align reduces over-refusal and improves robustness by decoupling spurious shortcuts in safety alignment of reasoning models.

  Zhejiang University, Ant Group · Sept 2026

• Big Techcs.CV

  One vision backbone can continually absorb new foundation model capabilities

  GRAFT continually extends a vision backbone by distilling from new teachers while using the previous model to retain old capabilities.

  University of Illinois Chicago, Amazon · Oct 2026

• Big Techcs.IR

  Co-Adaptive Routing and Retrieval Learns Query Rules Without Supervision

  CoRE jointly learns reasoning rule routing and constrained retrieval for structured knowledge without query-rule supervision.

  University of Georgia, Michigan State University · Oct 2026

• Industrycs.IR

  Shared embeddings let smaller query models search larger models’ document indexes

  A 0.6B and a 9B model distilled from one 18B teacher achieve state-of-the-art late-interaction retrieval and can be used asymmetrically.

  Perplexity AI · Oct 2026

• Research Labcs.AI

  Optimal transport aligns pretrained modalities with only hundreds of paired anchors

  Aligning two frozen unimodal encoders requires only 100 to 500 paired anchors when anchor-relative coordinates, entropic optimal transport, and distillation are combined.

  Sorbonne Université, Institut du Cerveau - Paris Brain Institute - ICM · Oct 2026

• Chinese Techcs.CL

  LLMs often diagnose without correctly valuing available medical evidence

  LLMs often submit correct diagnoses on insufficient or misleading evidence, and verifying evidential support before submission improves diagnostic accuracy.

  Zhejiang University, Ant Healthcare, Ant Group · Sept 2026

• Research Labcs.LG

  30 hours of fMRI per subject for decoding natural speech.

  Deep per-subject fMRI data improves naturalistic speech decoding performance continuously with more training data.

  Karavela, UNICOG, CNRS, INSERM, CEA, Paris-Saclay University · Oct 2026

• Chinese Techcs.CV

  CoVisco compresses long videos inside the vision encoder

  CoVisco uses abstract tokens and segmented attention to compress visual tokens, enabling 400-token video understanding that matches or exceeds OneVision-Encoder.

  ERNIE Team, Baidu Inc., The Hong Kong University of Science and Technology · Oct 2026

• Academiccs.AI

  Vision-language alignment fails to improve text-only spatial reasoning

  VQA-style vision-language alignment does not improve text-only reasoning on visualization tasks; self-distillation of visual embeddings into text does.

  New York University · Oct 2026

• Independentcs.IR

  Vector memory helps recommendation agents use full interaction histories

  CoVeMem replaces text-based user memory in agentic recommenders with graph-trained vector states, matching or beating text memory on 19 of 20 metric cells with zero per-interaction memory calls.

  Aug 2026