Papers for

content creators

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Large language model errors inspire creativity in different work stages

From LLM Hallucinations to Inspiration in Creative Work

Abstract: Large language models (LLMs) can introduce unexpected ideas through hallucinations, yet little is known about when creative practitioners find these errors inspiring or how these judgments shape interactions. We conducted a retrospective qualitative study with 20 experienced practitioners in research, design, and social media content creation, thematically analyzing 41 hallucination cases and 20 semi-structured interviews. Our framework links inspiration judgments of hallucinations based on goal relevance, task stage, content adaptability, and type-specific factors to behaviors across not-inspired, inspired-by, and inspired-to states. We characterize downstream interaction consequences, showing how hallucination-inspired experiences vary across creative stages and reshape role-differentiated trust, and propose these consequences as a dimension of hallucination evaluation. Finally, we derive design recommendations for stage-sensitive control of divergence and adherence, user-defined constraint priorities, differentiated hallucination labels, and salvage support for preserving useful ideas while correcting errors. These contributions deepen HCI understanding of hallucinations and responsible LLM design in creative work.

Mon 9 NovHuman-Computer Interaction
The gist
Sometimes, large language models (LLMs) make mistakes called hallucinations, where they create unexpected or wrong ideas. The authors studied how creative workers in research, design, and social media decide when these mistakes actually inspire them. They found that inspiration depends on how relevant the idea is, the task phase, and how easily the idea can be adapted. Their work also shows how this inspiration affects trust and interaction with the AI, and suggests ways to design AI tools that better handle these creative moments.
Open → 2611.12337v1

Prism improves language model output diversity and quality together

PRISM: Promoting Sampling Diversity of Large Language Models with Implicit Stochastic Modulation

Abstract: Large language models (LLMs) exhibit mode collapse, where their output distribution concentrates on familiar semantic modes. When diversity disappears, repeated sampling becomes semantic repetition rather than discovery, making rare but stronger solutions unlikely to surface and leaving self-evolution and continual learning difficult to sustain. Most remedies intervene only at inference time, shaping sam- ples without making diversity a model capability. We propose PRISM, a training-time framework that internalizes diversity within the model. PRISM combines (i) submodular curation that pairs each prompt with diverse-yet-correct multi-teacher solutions, and (ii) stochastic embedding modulation, which turns special-token perturbations into implicit signals for sampling distinct outputs. A theoretical analysis establishes the stability of this mechanism. On WritingPrompts and BookMIA, PRISM moves the quality-diversity frontier outward, achieving the best generation quality in all model-benchmark settings while maintaining the strongest or near-strongest semantic diversity across two open-weight model families. It also improves sampling efficiency on math and code benchmarks. Beyond inference, PRISM yields a stronger foundation for reinforcement learning, sustaining diverse rollouts and mitigating the diversity collapse commonly observed during RL training.

Fri 23 OctMachine Learning
The gist
Large language models often get stuck repeating familiar ideas instead of exploring new ones, which limits creativity and learning. The authors propose PRISM, a way of training models to produce more varied yet correct responses by showing them many diverse examples and tweaking certain parts of their processing to encourage variety. This approach helps the model balance quality with diversity, making it more likely to find rare but better answers and improving tasks like writing stories, math, and coding. PRISM also supports future learning steps by keeping output variety during reinforcement learning.
Open → 2610.32101v1

Stable and consistent multi-turn image editing with autoregressive diffusion

MT Forcing: Towards Stable and Consistent Multi-Turn Image Editing with Autoregressive Diffusion

Abstract: Multi-turn image editing aims to modify an image according to a sequence of instructions, enabling more effective interactive visual content creation for users. However, existing methods often exhibit noticeable drift after only a few editing turns, manifesting in two aspects: 1) unstable visual quality, with accumulating artifacts and degraded details; and 2) inconsistent image content, with unintended changes to preserved regions and prior edits. To address these challenges, we propose MT Forcing, a novel training paradigm for multi-turn image editing with unified multimodal models (UMMs) based on autoregressive diffusion. Specifically, we propose Single-Turn Distribution Matching Distillation (Single-Turn DMD), which supervises student self-rollouts with a teacher that applies cumulative edits to the source image in a single turn, providing visual-quality guidance without recursive error accumulation. Additionally, we introduce Teacher-Forcing Guidance, which supervises history-dependent transitions on reference trajectories during warmup and as a regularizer during distillation, encouraging cross-turn content consistency beyond what per-turn distribution matching alone can ensure. Extensive experiments demonstrate that our approach significantly improves the multi-turn image editing capabilities of UMMs, yielding more stable visual quality and stronger cross-turn consistency.

Sun 18 OctComputer Vision and Pattern Recognition
The gist
Editing an image multiple times in a row often causes it to lose quality or change in unintended ways. The authors developed a new training method called MT Forcing to help AI models edit images step-by-step without making mistakes that pile up. Their approach trains the model to keep the image stable and consistent across many editing turns. Tests show it improves the quality and accuracy of multi-step image edits.
Open → 2610.25404v1

Semantic critical visual regions improve video to audio generation

Semantic-Critical Regions for Video-to-Audio Generation

Abstract: In real-world videos, sound events are often associated with localized and fine-grained visual cues, such as object motion, object-part dynamics, contact events, and interactions between objects. Existing video-to-audio (V2A) methods employ implicit conditioning based on global visual representations or explicit control derived from temporal and spatial cues. However, subtle sound-producing and sound-indicative interactions remain insufficiently captured, leaving room for more precise visual conditioning of sound-event semantics and temporal dynamics. Motivated by this observation, we refer to regions containing such sound-relevant visual evidence as Semantic-Critical Regions (SCRs) and introduce SCR-V2A, a V2A generation framework that explicitly localizes and encodes SCRs as fine-grained visual guidance for audio generation. Specifically, we first identify SCRs through sound-related visual localization. We then transform the localized regions into a control representation through adapter-based dual visual encoding, which jointly captures global visual context and fine-grained sound-relevant evidence. The representation is used to condition a downstream audio generation model. Experiments on both in-domain and out-of-domain benchmarks demonstrate that SCR-V2A achieves strong performance in distribution matching and audio quality while remaining competitive in cross-modal alignment and requiring few trainable parameters. These results highlight the value of prioritizing fine-grained, sound-relevant visual regions for V2A generation.

Fri 16 OctComputer Vision and Pattern RecognitionMultimediaSound
The gist
Sometimes, the sounds in videos come from small, specific parts of the scene, like moving objects or interactions. The paper introduces a way to find these important spots, called Semantic-Critical Regions (SCRs), and uses them to guide the creation of matching sounds. The authors showed that focusing on these regions helps produce better and clearer audio from videos. Their method also works well on different kinds of videos and doesn't require training many new parameters.
Open → 2610.23021v1

Training stages change large language model ability to persuade people

Successive Training Stages and Large Language Model Persuasion: Effects of Misalignment, Supervised Fine-Tuning, and Preference Optimization

Abstract: Large language models (LLMs) can be tuned to influence human attitudes, yet the respective contributions of successive post-training stages remain un-clear. This study examines how three successive training stages affect LLM persuasiveness: (1) misalignment through supervised fine-tuning (SFT) on conspiracy data, (2) additional persuasive SFT on argumentative data, and (3) Identity Preference Optimization (IPO), a preference-optimization method. A total of 835 participants recruited on Prolific were randomly assigned to five between-subject conditions (neutral text, conspiracy-trained model, persuasion-trained model, preference-optimized model, and GPT-4) and were exposed to texts on 10 divisive political issues, personalized from their individual profiles in all model conditions. Attitude change was measured as the difference between pre- and post-exposure positions on continuous Likert scales and analyzed with an analysis of covariance (ANCOVA). A significant condition x baseline-attitude interaction, F (4, 825) = 5.33, p < .001, indicated that training effects depended on participants' initial attitudes. Persuasive SFT produced greater attitude change than conspiracy training alone, d = 0.30, whereas IPO provided no additional benefit, d = 0.03, and GPT-4 did not differ from neutral text, d = --0.01. These results show that targeted supervised training on persuasive data increases LLM persuasiveness, whereas preference optimization yields no significant gains beyond it.

Wed 7 OctArtificial IntelligenceComputers and Society
The gist
Large language models can be trained in different ways to try to persuade people changing their opinions. This study tested models trained on conspiracy ideas, on arguing persuasively, and on optimizing preferences to see which influenced attitudes more. The models trained specifically on persuasive arguments changed people’s attitudes more than the conspiracy-trained model, but the preference optimization did not help much. The standard GPT-4 model showed no difference from neutral text. This shows focused training on persuasive data matters for changing beliefs.
Open → 2610.09964v1

DimSteer helps writers control AI outputs with intuitive style sliders

DimSteer: Steering LLM Authoring with Automatically Discovered Stylistic Controls

Abstract: Large language model writing interfaces often make users steer outputs by repeatedly articulating desired changes in natural language. Yet writers may recognize useful stylistic directions only after seeing alternatives, making revision recall-heavy. We present DimSteer, an authoring interface that samples prompt-local completions, discovers high-variance activation-space axes of variation, labels them, and exposes them as sliders with pole previews, diff comparison, and reset controls. Users can manipulate discovered dimensions, reducing the need to reformulate prompts for each stylistic adjustment. In a within-subjects study with 16 participants against a matched prompt-only baseline, DimSteer reduced mental demand, effort, and frustration while preserving comparable perceived success. Participants valued the surfaced dimensions, yet 15 of 16 disagreed that they would have thought to request the same changes in a prompt. Results suggest prompt-local controls can shift LLM authoring from recall-based prompting toward recognition-based exploration and direct manipulation, while preserving prompting for open-ended edits.

Sat 3 OctHuman-Computer InteractionComputation and Language
The gist
When using AI to write, it can be hard to get the style just right by typing instructions alone. The authors created DimSteer, a tool that shows users different writing versions and finds key style directions automatically. Users can then adjust these styles easily with sliders instead of retyping prompts. This method reduces the effort and frustration of editing while still letting users explore new writing styles.
Open → 2610.04174v1

Memory guided system creates b roll sequences from user videos

Memory-Guided B-Roll Generation from User Video Collections

Abstract: We introduce an approach for collection-grounded B-roll sequence generation. Given a user's video collection, a directive given in natural language, and a target duration, the goal is to produce a multi-shot sequence that complements the user's primary footage (A-roll) while preserving the collection's characters, settings, objects, and style. This task is challenging as one must choose the visual evidence from hours of captured footage that should guide the generation of each shot in the sequence. We address this challenge with MemComposer, a three-stage system that turns raw footage into a structured memory with visual references (characters, settings, objects, and style) and uses it to plan, retrieve, and generate grounded B-roll sequences. First, in a one-time offline stage, MemComposer constructs an entity-centric memory from raw video. Second, it uses the memory and user directive to plan a grounded sequence and retrieve conditioning frames for each shot. Third, it iteratively generates and critiques the sequence to enforce identity, setting, and sequence-level consistency. We evaluate MemComposer in a user preference study along two dimensions: prompt adherence and visual alignment to the user's collection. Against an ungrounded text-to-video planner, MemComposer wins 60.0\% of prompt-adherence and 92.8\% of visual-alignment comparisons, showing the grounding benefit of collection memory and reference retrieval. Against retrieval-only sequences assembled from captured footage, MemComposer wins 94.5\% of prompt-adherence comparisons, showing the value of generating missing shots, while retrieval-only sequences are preferred for visual alignment in 58.2\% of comparisons.

Thu 1 OctComputer Vision and Pattern Recognition
The gist
Editing videos with interesting extra footage is hard when you have hours of clips to choose from. The authors introduce MemComposer, a tool that learns what’s in a user’s video collection and then plans and generates new video sequences that match the people, places, and style seen before. It picks the best frames and makes sure the new sequences fit what the user wants and look consistent. Their tests show MemComposer produces better matching and more relevant sequences than other methods that don’t use this memory or only reuse existing clips.
Open → 2610.01884v1

Tacit-TTS speeds up voice cloning without using transcripts

Tacit-TTS: From Autoregressive Decoding to Masked Prediction for Efficient Transcript-Free Voice Cloning

Abstract: TTS systems with autoregressive semantic modeling have demonstrated strong zero-shot voice cloning performance and rich expressive variation, but their sequential decoding incurs substantial latency. Non-autoregressive alternatives offer much faster generation, yet often rely on more restrictive reference conditioning, such as requiring transcripts of the reference speech during inference. We present Tacit-TTS, an efficient transcript-free zero-shot voice cloning system distilled from IndexTTS2. Our model replaces autoregressive text-to-semantic decoding with masked non-autoregressive generation, introduces training-free acoustic length estimation, and accelerates the flow-matching renderer through ReFlow distillation. Across two English and two Mandarin datasets, Tacit-TTS achieves competitive zero-shot quality while generating speech over 10x faster than IndexTTS2 for utterances longer than 5 seconds. Its transcript-free conditioning further supports cross-lingual and non-lexical references. We validate this capability using references from eight other languages, infant babble, and synthetic gibberish, where transcript-dependent systems often degrade or fail due to unreliable ASR transcripts.

Tue 29 SeptComputation and LanguageSound
The gist
Voice cloning systems can copy a person's voice by analyzing speech, but many need to process words one after another, which is slow, or require the text of what was said, which limits flexibility. The authors created Tacit-TTS, a system that copies voices quickly without needing the transcript of the original speech. This means it works well across different languages and even with sounds like baby talk or made-up words. Tacit-TTS is much faster than earlier methods while keeping similar quality in cloned voices.
Open → 2609.38658v1

Trigger indexed memory improves personalized text to image generation

V-Engram: Trigger-Indexed External Memory for Modular Text-to-Image Personalization

Abstract: Pretrained text-to-image models contain broad visual knowledge, yet they cannot reliably acquire or refine a specific visual identity from only a few references while preserving compositional control. Token-embedding methods are compact but often underfit identity, whereas adapter-based methods improve fidelity through persistent weight updates that can be costly to store and interfere when concepts are composed. We introduce V-Engram, a trigger-indexed external memory mechanism for Stable Diffusion 3.5. Each concept is assigned an explicit trigger that retrieves concept-specific memory, whose gated directions enter frozen text-encoder and MMDiT context states as relative residuals. Separating this memory from backbone adaptation enables prompt-selective and multi-concept access without merging model updates. Experiments show that V-Engram broadly matches DreamBooth-LoRA in overall subject fidelity while showing advantages in settings such as contextual subject preservation. Prompt-matched loading retrieves only matched entries, reducing most additional adaptation-state loading for a single-concept query. Qualitative results further demonstrate paired-trigger composition and same-class separation, while prompts without registered entries retain the frozen model's base behavior. Together, these results establish trigger-indexed memory as a modular interface for adding targeted visual evidence without rewriting the generator.

Tue 29 SeptArtificial Intelligence
The gist
Pretrained text-to-image models are good at generating images from descriptions but struggle to accurately create specific characters or objects from just a few examples. The authors present V-Engram, a method that stores detailed visual info separately and links it to special keywords or 'triggers.' This helps the model remember and use specific appearances without changing its main knowledge, allowing for multiple personalized items in one prompt and better control. Their tests show V-Engram matches other methods in image accuracy while being more flexible and efficient.
Open → 2609.37198v1

Framework makes human and AI understandings match better in tasks

Alignment Games: A Framework for Conceptual Repair in Human-AI Collaboration

Abstract: The meaning of a concept in use is shaped by the situation, task, goals, and prior knowledge. For example, a request to make a poster "visually appealing for a five-year-old" might evoke bright colors and cartoon imagery for one collaborator, but less text, bold shapes, and visual simplicity for another. We call such task-relevant differences conceptual misalignment. We introduce Alignment Games, a framework for making these differences visible and repairable during human-AI interaction. Drawing on theories of situated conceptualization, we characterize task-specific conceptual frames in terms of relevant attributes, values, relations, constraints, and priorities. We then define alignment moves that intervene on the situation, the reasoning used to interpret it, or the resulting frame. Through examples from educational content generation, creative coding, and argumentative writing, we show how these moves can be composed into repair sequences and derive design principles for supporting task-sufficient conceptual alignment at runtime.

Mon 28 SeptHuman-Computer InteractionArtificial Intelligence
The gist
People can have different ideas about what a task means, which causes confusion when working with AI. The authors introduce a method called Alignment Games that helps show and fix these differences while humans and AI collaborate. This method looks at what matters in a task, like priorities and rules, and allows changing either the situation or the way the AI thinks to reach better agreement. They demonstrate how this approach helps in creating educational materials, code, and arguments. This makes it easier for AI tools to understand what people really want.
Open → 2609.35197v1

Diffusion transformers enhance real-world image super-resolution efficiently

Fill2SR: Repurposing Inpainting Diffusion Transformers for Real-World Super-Resolution

Abstract: Recent real-world image super-resolution (SR) methods often adapt text-to-image (T2I) backbones with ControlNet-style branches or spatial conditioning tokens, which increases memory and computes with resolution and often constrains training to a fixed scale. We propose Fill2SR, which repurposes a masked-inpainting Diffusion Transformer for SR without extra spatial branches. Our Inpainting-Interface Evidence Adapter (IIEA) writes the low-quality (LQ) observation into the native masked-image slot under a full-image mask, turning inpainting into a reverse-degradation conditional rectified flow trained with LoRA-only tuning. We further introduce RCDT, an offline pipeline that distills degradation descriptors from unpaired real images and transfers them onto clean targets using frozen open-source models. Fill2SR supports mixed-resolution training up to QHD and yields stable performance across $512/1024/2048$ outputs. On synthetic benchmarks, our base model with IIEA achieves the best LPIPS on DIV2K and LSDIR; adding RCDT trades a small LPIPS drop for consistently stronger no-reference quality on RealLQ250 and RealPhoto60. Fill2SR remains memory-predictable, running $1536^2$ inference on a single 32GB GPU and extending to multi-megapixel outputs via tiled restoration.

Sun 27 SeptComputer Vision and Pattern Recognition
The gist
Improving low-quality images to higher resolutions is usually costly in memory and limited by fixed scales. The authors introduce Fill2SR, which cleverly reuses an image-inpainting diffusion transformer to enhance images without extra memory-heavy parts. They also develop a way to learn about real image degradations and apply that knowledge to improve clean images. Their method works well on different image sizes and runs efficiently on common hardware.
Open → 2609.33582v1

Cognitive multi-agent system creates persuasive videos like humans

CogenPVG: Cognitive-Enhanced Reflective Multi-Agent Framework for Persuasive Video Generation

Abstract: Persuasive video generation (PVG) is a valuable yet under-explored research topic. Despite the significant advances in multimodal content generation, AI-empowered automated creation of human-made-like videos with substantial persuasiveness remains a formidable challenge. In this paper, we propose CogenPVG, a novel Cognitive-Enhanced reflective multi-agent framework tailored for Persuasive Video Generation task. Given the topic and stance from the user, we decouple the sophisticated generation process into four sequential stages: argument reasoning, storyboard planning, asset creation, and post-editing, imitating the workflow of human video producers. To ensure high persuasiveness, each stage is equipped with a pair of generator and critic agents, following a reflective refinement scheme grounded in a solid psychological theory of persuasion, the Elaboration Likelihood Model (ELM). In the argument reasoning stage, we generate highly logical and credible reasoning thoughts under the guidance of critical thinking theory, enabling cognitive enhancement via the central route of the ELM. For the other three stages, we generate and optimize multimodal assets, assembling them into a persuasive video guided by theories of heuristics, as the peripheral route of the ELM. To the best of our knowledge, CogenPVG is the first work focused on general persuasive topics, without being confined to commercial purposes. Extensive experiments and comprehensive analysis demonstrate that our framework achieves the best persuasion performance, thereby proving the effectiveness of our proposed multi-agent framework for the PVG task.

Tue 22 SeptMultimediaArtificial Intelligence
The gist
Making videos that can persuade people is hard for computers. The authors created a system called CogenPVG that works like a team of agents, each doing a step like thinking of arguments, planning scenes, making content, and editing. These agents check and improve each other's work using ideas from psychology about how people get convinced. This approach makes videos that feel logical and appealing, much like those made by humans.
Open → 2609.25821v1

Netflix reveals new way to measure recommendation impact accurately

Beyond Raw Engagement: A Counterfactual Observability Framework for Recommender Systems at Netflix

Abstract: Understanding the performance of large-scale recommender systems remains an underexplored challenge, especially for content creators and model developers. The raw engagement signals available to them, such as views and clicks, conflate content quality, model behavior, presentation bias, and audience reach, making it hard to attribute outcomes to the right cause. In this work, we present a general evaluation framework that enhances observability across multiple recommender systems at Netflix and demonstrate its effectiveness through several production deployments. The framework treats recommender-system observability as a counterfactual measurement problem: estimating what the recommender would have done, and what engagement would have followed, in the absence of a specific content item or model decision. We articulate three stakeholder-centered observability principles for content creators and model developers, and propose measurement methodologies covering bias reduction, relativity, and incrementality, applicable to both single-stage and cascading recommender systems and serving both audiences from a single measurement foundation.

Sat 19 SeptInformation RetrievalArtificial Intelligence
The gist
Netflix has a new way to figure out how well its recommendation system really works. Often, just looking at clicks or views doesn’t tell the full story because many things can influence those numbers. The authors created a method that estimates what would have happened if a certain show or algorithm choice hadn't been made. This helps creators and engineers understand the real reasons behind user engagement more clearly. It works across different recommendation stages and helps reduce bias in measuring results.
Open → 2609.22747v1

Ownership matters when people lead AI-assisted everyday tasks

Ownership in AI-Assisted Everyday Tasks

Abstract: When does work done with AI still feel like ours? As AI becomes woven into everyday tasks, we must examine what happens to our sense of ownership and contribution when a machine shares in producing what we make. We report an exploratory qualitative survey in which participants were asked to describe two recent, self-selected tasks completed with AI: one that felt like their own and one that did not. We find that felt ownership depends on the process of collaboration: people disown work when they merely approve AI's suggestions, but retain ownership when they lead, iterate, or rewrite. Ownership can also extend to settings where people own the vision for a project but not the execution; respondents reported high ownership on tasks they could not have completed without AI. Loss of personal voice and a lack of comprehension of the output both erode ownership. Finally, willingness to disclose AI use is often decoupled from actual pride or ownership, and instead shaped by community norms and fear of credit erasure. We propose several research directions as a result of these findings to promote AI development that supports people's sense of authorship over their own lives.

Thu 17 SeptArtificial IntelligenceHuman-Computer Interaction
The gist
This paper looks at when people still feel like the work they do with AI is really theirs. The authors find that people feel ownership when they actively guide or change the AI’s work, but less so when they just agree with what the AI suggests. People also keep ownership when they have the original idea, even if AI does much of the work. Losing a personal touch or not understanding the AI output can make people feel less connected to the work. Finally, whether people tell others they used AI often depends on social norms or worries about losing credit, not just pride in the work itself.
Open → 2609.20658v1