{"data":{"event":{"id":"5b63746e-d332-468c-a495-1b383c525542","slug":"converge-then-diversify-decoupling-convergence-and-diversity-in-multi-ob-4f5159b6ce","title":"Converge Then Diversify: Decoupling Convergence and Diversity in Multi-Objective Bayesian Optimisation","short_summary":"arXiv:2609.13396v1 Announce Type: new \nAbstract: Multi-objective Bayesian optimisation (MOBO) is a sample-efficient approach for optimising expensive black-box functions with multiple objectives. In MOBO, the goal is to adequately approximate the Pareto front; that is, to obtain a high-quality solution set with 1) good convergence (closeness to the Pareto front) and 2) good diversity (spread across the Pareto front). Existing MOBO methods typically aim to accomplish these two tasks simultaneously, i.e., driving the search towards the Pareto front while maintaining a diverse set of nondominated","full_description":"arXiv:2609.13396v1 Announce Type: new \nAbstract: Multi-objective Bayesian optimisation (MOBO) is a sample-efficient approach for optimising expensive black-box functions with multiple objectives. In MOBO, the goal is to adequately approximate the Pareto front; that is, to obtain a high-quality solution set with 1) good convergence (closeness to the Pareto front) and 2) good diversity (spread across the Pareto front). Existing MOBO methods typically aim to accomplish these two tasks simultaneously, i.e., driving the search towards the Pareto front while maintaining a diverse set of nondominated solutions, such that the solutions, ideally, can gradually approach the entire front. When sufficient search budgets are available, this approach is effective. However, considering both convergence and diversity throughout the search is not easy and requires careful design. Under very tight budgets, there may not be enough solutions generated to be able to simultaneously approach the entire Pareto front. To address this issue, this paper proposes a \\textit{converge-then-diversify} (CTD) approach that decouples convergence and diversity into two stages. In the first stage, CTD focuses on convergence, aiming to quickly drive the search toward a single point on the Pareto front. In the second stage, CTD focuses on diversity, aiming to spread solutions across the front. We present two simple instantiations of CTD by using widely adopted acquisition functions in the area. Experimental results show that, across all 446 pairwise comparisons, CTD statistically outperforms state-of-the-art methods in 72.9\\% of the cases, performs equivalently in 21.1\\%, and is statistically worse in only 6.1\\%, with the advantage being particularly evident in settings with very tight evaluation budgets or in high-dimensional problems.\n\nUpdate (2026-09-17): Towards training-time mitigations for alignment faking in RL We construct a diverse array of model organisms of alignment faking and study mitigations that could be used during RL to decrease alignment faking and compliance gaps.. Towards training-time mitigations for alignment faking in RL We construct a diverse array of model organisms of alignment faking and study mitigations that could be used during RL to decrease alignment faking and compliance gaps.\n\nUpdate (2026-09-18): LLM-as-an-Improver: Turning Verification into Better Candidates. arXiv:2609.19515v1 Announce Type: new \nAbstract: Verifier-based selection improves LLM performance by generating multiple candidate solutions and using a verifier to select the most promising one. However, existing methods typically treat verification only as a ranking step and discard its feedback once a fixed candidate pool has been evaluated. In this paper, we ask whether verification can also improve the candidate set itself. To this end, we introduce LLM-as-an-Improver and propose Verify--Repair--Reselect (VRR), which uses verification feedback to generate and reselect improved candidates. VRR retains the initial winner while conditionally generating three complementary alternatives: repaired versions of the winner and runner-up, and a solution based on a new approach. It filters invalid and duplicate candidates using only inference-time information and then reselects the final answer under the original evaluation criteria. Across diverse models and code-generation and reasoning benchmarks, VRR improves over fixed-pool verifier-based selection in many settings and can recover correct soluti","ledger_type":"benefit","primary_domain_id":"8f1af1b9-7302-4b51-aae9-fdfac04d158a","event_status":"provisional","event_date":"2026-09-15T00:00:00.000Z","discovery_date":"2026-09-15T00:00:00.000Z","first_published_date":"2026-09-15T00:00:00.000Z","last_reviewed_date":"2026-09-18T00:00:00.000Z","geographic_scope":"International","affected_population":null,"base_impact_tier":1,"base_score":"1.00","attribution_multiplier":"0.1000","evidence_multiplier":"0.3000","realization_multiplier":"0.2000","durability_multiplier":"0.5000","current_event_score":"0.003000","confidence_level":"low","score_explanation":"Ingest update: independent source raised evidence; added a follow-up source. Score remains provisional until review. Follow-up ingest: independent source raised evidence; added a follow-up source.","methodology_version_id":"8745541b-0f8e-4d5a-9b47-5291d3813a8d","original_methodology_version_id":"8745541b-0f8e-4d5a-9b47-5291d3813a8d","published_at":"2026-09-15T04:00:54.685Z","created_at":"2026-09-15T04:00:54.685Z","updated_at":"2026-09-18T04:00:56.453Z","flags":[],"domain_name":"Biology","domain_slug":"biology","methodology_version":"0.2"},"contributions":[{"id":"2e5a150f-188b-4b96-9723-090397afebe3","event_id":"5b63746e-d332-468c-a495-1b383c525542","model_id":"535014f7-c941-4d39-aa9b-8f9c1d3d07b7","role_description":"Unspecified system mentioned or implied by a news item. Remap to a named release when identified.","attribution_multiplier":"0.1000","credit_share":"1.0000","contribution_score":"0.003000","attribution_rationale":"News ingest does not infer a named model from the publisher alone. Attribution stays unspecified until a release is identified.","attribution_confidence":"medium","first_used_date":"2026-09-15T00:00:00.000Z","model_version_if_known":null,"review_status":"approved","created_at":"2026-09-15T04:00:54.695Z","model_slug":"unspecified-ai-system","model_name":"Unspecified AI system","identity_class":"unknown","is_internal":false,"family_name":"Unspecified","family_slug":"unknown-unspecified","organization_name":"Unknown","organization_slug":"unknown"}],"sources":[{"id":"7cf5ffd7-522f-4f11-af91-bd5b5f4ad930","event_id":"5b63746e-d332-468c-a495-1b383c525542","url":"https://arxiv.org/abs/2609.13396","canonical_url":"https://arxiv.org/abs/2609.13396","source_type":"preprint","publisher":"arXiv cs.AI","author":null,"publication_date":"2026-09-15T00:00:00.000Z","retrieved_at":"2026-09-15T04:00:54.701Z","title":"Converge Then Diversify: Decoupling Convergence and Diversity in Multi-Objective Bayesian Optimisation","excerpt":"arXiv:2609.13396v1 Announce Type: new \nAbstract: Multi-objective Bayesian optimisation (MOBO) is a sample-efficient approach for optimising expensive black-box functions with multiple objectives. In MOBO, the goal is to adequately approximate the Pareto front; that is, to obtain a high-quality solution set with 1) good convergence (closeness to the Pareto front) and 2) good diversity (spread across the Pareto front). Existing MOBO methods typically aim to accomplish these two tasks simultaneousl","content_hash":null,"source_reliability_class":"medium","is_primary_source":true,"is_independent":true,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-15T04:00:54.701Z"},{"id":"857efd3e-00fa-46c1-acef-e546a4001545","event_id":"5b63746e-d332-468c-a495-1b383c525542","url":"https://alignment.anthropic.com/2025/alignment-faking-mitigations/","canonical_url":"https://alignment.anthropic.com/2025/alignment-faking-mitigations/","source_type":"developer_announcement","publisher":"Anthropic Alignment Science","author":null,"publication_date":"2026-09-17T00:00:00.000Z","retrieved_at":"2026-09-17T03:01:11.333Z","title":"Towards training-time mitigations for alignment faking in RL We construct a diverse array of model organisms of alignment faking and study mitigations that could be used during RL to decrease alignment faking and compliance gaps.","excerpt":"Towards training-time mitigations for alignment faking in RL We construct a diverse array of model organisms of alignment faking and study mitigations that could be used during RL to decrease alignment faking and compliance gaps.","content_hash":null,"source_reliability_class":"medium","is_primary_source":false,"is_independent":true,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-17T03:01:11.333Z"},{"id":"83099c71-0ffb-4521-bb82-bca567203963","event_id":"5b63746e-d332-468c-a495-1b383c525542","url":"https://arxiv.org/abs/2609.19515","canonical_url":"https://arxiv.org/abs/2609.19515","source_type":"preprint","publisher":"arXiv cs.AI","author":null,"publication_date":"2026-09-18T00:00:00.000Z","retrieved_at":"2026-09-18T04:00:56.438Z","title":"LLM-as-an-Improver: Turning Verification into Better Candidates","excerpt":"arXiv:2609.19515v1 Announce Type: new \nAbstract: Verifier-based selection improves LLM performance by generating multiple candidate solutions and using a verifier to select the most promising one. However, existing methods typically treat verification only as a ranking step and discard its feedback once a fixed candidate pool has been evaluated. In this paper, we ask whether verification can also improve the candidate set itself. To this end, we introduce LLM-as-an-Improver and propose Verify--R","content_hash":null,"source_reliability_class":"medium","is_primary_source":false,"is_independent":true,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-18T04:00:56.438Z"}],"claims":[{"id":"aacff661-bd57-4904-bd10-61466ba5f91f","event_id":"5b63746e-d332-468c-a495-1b383c525542","claim_text":"Converge Then Diversify: Decoupling Convergence and Diversity in Multi-Objective Bayesian Optimisation","claim_type":"outcome","claim_status":"supported","confidence_score":null,"created_at":"2026-09-15T04:00:54.707Z","updated_at":"2026-09-15T04:00:54.707Z"}],"revisions":[{"id":"d208cfc4-eb23-49b9-90df-b6592f9e087f","event_id":"5b63746e-d332-468c-a495-1b383c525542","model_id":null,"previous_score":"0.000000","new_score":"0.001000","previous_factors":{},"new_factors":{"evidence":0.1,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"change_reason":"Auto-published from news ingest.","trigger_type":"news_ingest","trigger_source_ids":null,"reviewer_id":null,"review_status":"published","created_at":"2026-09-15T04:00:54.718Z"},{"id":"72e6d85f-ee0e-49a0-bf24-e46508a1baf3","event_id":"5b63746e-d332-468c-a495-1b383c525542","model_id":null,"previous_score":"0.001000","new_score":"0.002000","previous_factors":{"tier":1,"evidence":0.1,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"new_factors":{"tier":1,"models":["535014f7-c941-4d39-aa9b-8f9c1d3d07b7"],"evidence":0.2,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"change_reason":"Ingest update: independent source raised evidence; added a follow-up source.","trigger_type":"ingest_update","trigger_source_ids":["857efd3e-00fa-46c1-acef-e546a4001545"],"reviewer_id":null,"review_status":"published","created_at":"2026-09-17T03:01:11.353Z"},{"id":"b798bd53-fdf5-4655-92cf-36e14b11c8a0","event_id":"5b63746e-d332-468c-a495-1b383c525542","model_id":null,"previous_score":"0.002000","new_score":"0.003000","previous_factors":{"tier":1,"evidence":0.2,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"new_factors":{"tier":1,"models":["535014f7-c941-4d39-aa9b-8f9c1d3d07b7"],"evidence":0.3,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"change_reason":"Ingest update: independent source raised evidence; added a follow-up source.","trigger_type":"ingest_update","trigger_source_ids":["83099c71-0ffb-4521-bb82-bca567203963"],"reviewer_id":null,"review_status":"published","created_at":"2026-09-18T04:00:56.462Z"}],"secondary":[]},"methodology_version":"0.1"}