Search Results (3)

In long-horizon multi-party conversations, human-centric AI agents face a persistent structural problem: similarity-based retrieval may fail to reconnect semantically dispersed fragments of the same evolving event. This problem severely weakens causal continuity and multi-hop context recovery. To improve attribution trust and reduce structural erasure, we propose MemLoom, a dual-loop causality-traceable retrieval framework that organizes conversational history as an event memory graph. MemLoom decouples latency-sensitive online interaction from off-peak structural curation through online event formation, sentence-level buffering, asynchronous neuro-symbolic graph synthesis, and bounded dual-stream retrieval. Evaluations across QMSum, LoCoMo, and the synthetic causal diagnostic suite (SCDS) support the structural utility of MemLoom. For LoCoMo, under our unified local evaluation setup, MemLoom shows favorable temporal and multi-hop reasoning results (J = 65.77 and 58.14) relative to contemporary agentic baselines, such as Mem0, Zep, and A-Mem. For SCDS, within a controlled diagnostic setting, it recovers demanded causal chains more reliably than GraphRAG (SCR = 0.72 vs. 0.35) and maintains stronger answer-level auditability (AA = 0.80 vs. 0.50). This is achieved with a bounded online P95 latency of 1.67 s. These results indicate that asynchronous dual-loop stewardship has practical value for causality-traceable, event-centric conversational memory in multi-party settings. Full article

Against the dual backdrop of rapid population aging and legacy neighborhood renewal, morphologically planning-locked legacy neighborhoods in high-density cities face persistent imbalances in outdoor activity spaces that undermine aging-in-place participation and health equity. This study advances a Spatial Pathology framework. Using nine representative communities in Longtan Subdistrict, Dongcheng District, Beijing, we develop a GIS-assisted spatial audit, a systematic behavioral observation protocol with temporal-intensity metrics, and a validated perception instrument. These tools form a closed evidentiary loop with explicit indicator definitions, formulas, and decision thresholds, alongside a reproducible analytic and visualization pipeline. Tri-dimensional baselines revealed substantial inter-community disparities: Spatial Quality Index (SQI) ranged from 43.3 to 77.0; activity intensity varied from 1.5 to 15.7 persons/100 m²·hour; and overall satisfaction scores spanned 3.88–4.49. It quantifies and identifies three core paradoxes in outdoor activity spaces within this context: (1) the Functional Failure Paradox with FFI exceeding +0.5 and ELR surpassing 60% in dormant communities; (2) the Value Misalignment Paradox where Facilities & Equipment showed the strongest satisfaction impact (β = 0.344) yet the largest unmet-need gap (VQGI > +8); (3) the Practice–Perception Decoupling Paradox evidenced by a negative correlation (r = −0.38) between usage intensity and satisfaction. These paradoxes reveal the spatial roots of planning-locked legacy neighborhoods—compound mechanisms of planning inertia, decision–demand information gaps, and elderly adaptability masking environmental deficits. We translate the diagnosis into typology-specific prescriptions—reactivating dormant spaces via “route–node–plane” continuity and proximal micro-spaces; decongesting peak periods through elastic zoning and equipment redistribution; and precision calibration of facilities and walking loops—implemented through co-creation and light-touch stewardship. This provides evidence-based, precision-targeted intervention pathways for micro-renewal of aging neighborhoods, supporting localized implementation of UN Sustainable Development Goals (SDG 11 Sustainable Cities; SDG 10 Reduced Inequalities). This methodological framework is transferable to other high-density aging cities, offering theoretical scaffolding and empirical reference for multi-source geographic data-driven urban spatial analysis and equity-oriented age-friendly retrofitting. Full article

In less than 30 years, Deep Brain Stimulation (DBS) has evolved from an antiparkinsonian rescue intervention into a flexible neuromodulatory therapy with the potential for personalized, adaptive, and enhancement-focused interventions. In this review we collected evidence from seven areas: (i) modern eligibility criteria, and ways to practically improve on these, outside of ‘Core Assessment Program of Surgical Interventional Therapies in Parkinson’s Disease’ (CAPSIT-PD); (ii) cost-effectiveness, where long-horizon models now show positive incremental net monetary benefit for Parkinson’s disease, and rechargeable-devices lead the way in treatment-resistant depression and obsessive–compulsive disorder; (iii) anatomical targets, from canonical subthalamic nucleus (STN) / globus pallidus internus (GPi) sites, to new dual-node and cortical targets; (iv) mechanistic theories from informational lesions, antidromic cortical drive, and state-dependent network modulation made possible by optogenetics and computational modeling; (v) psychiatric and metabolic indications, and early successes in subcallosal and nucleus-accumbens stimulation for depression, obsessive–compulsive disorder (OCD), anorexia nervosa, and schizophrenia; (vi) procedure- and hardware-related safety, summarized through five reviews, showing that the risks were around 4% for infection, 4–5% for revision surgery, 3% for lead malposition or fracture, and 2% for intracranial hemorrhage; and (vii) future directions in connectomics, closed-loop sensing, and explainable machine learning pipelines, which may change patient selection, programming, and long-term stewardship. Overall, the DBS is entering a “third wave” focused on a better understanding of neural circuits, the integration of AI-based adaptive technologies, and an emphasis on cost-effectiveness, in order to extend the benefits of DBS beyond the treatment of movement disorders, while remaining sustainable for healthcare systems. Full article