In situ concentration of semi-volatile aerosol using water-condensation technology

The effect of concentrating semi-volatile aerosols using a water-condensation technology was investigated using the Versatile Aerosol Concentration Enrichment System (VACES) and the Aerodyne Aerosol Mass Spectrometer (AMS) during measurements of ambient aerosol in Pittsburgh, PA. It was found that the shape of the sulfate mass-weighed size distribution was approximately preserved during passage through the concentrator for all the experiments performed, with a mass enhancement factor of about 10-20 depending on the experiment. The size distributions of organics, ammonium and nitrate were preserved on a relatively clean day (sulfate concentration around 7μg/m3), while during more polluted conditions the concentration of these compounds, especially nitrate, was increased at small sizes after passage through the concentrator. The amount of the extra material, however, is rather small in these experiments: between 2.4% and 7.5% of the final concentrated PM mass is due to "artifact" condensation. An analysis of thermodynamic processes in the concentrator indicates that the extra particle material detected can be explained by redistribution of gas-phase material to the aerosol phase in the concentrator. The analysis shows that the condensation of extra material is expected to be larger for water-soluble semi-volatile material, such as nitrate, which agrees with the observations. The analysis also shows that artifact formation of nitrate will be more pronounced in ammonia-limited conditions and virtually undetectable in ammonia-rich conditions. © 2004 Elsevier Ltd. All rights reserved.

Nested dictionary learning for hierarchical organization of imagery and text

A tree-based dictionary learning model is developed for joint analysis of imagery and associated text. The dictionary learning may be applied directly to the imagery from patches, or to general feature vectors extracted from patches or superpixels (using any existing method for image feature extraction). Each image is associated with a path through the tree (from root to a leaf), and each of the multiple patches in a given image is associated with one node in that path. Nodes near the tree root are shared between multiple paths, representing image characteristics that are common among different types of images. Moving toward the leaves, nodes become specialized, representing details in image classes. If available, words (text) are also jointly modeled, with a path-dependent probability over words. The tree structure is inferred via a nested Dirichlet process, and a retrospective stick-breaking sampler is used to infer the tree depth and width.

Inner speech and bilingual autobiographical memory: a Polish-Danish cross-cultural study.

Thirty years after fleeing from Poland to Denmark, 20 immigrants were enlisted in a study of bilingual autobiographical memory. Ten "early immigrators" averaged 24 years old at the time of immigration, and ten "late immigrators" averaged 34 years old at immigration. Although all 20 had spent 30 years in Denmark, early immigrators reported more current inner speech behaviours in Danish, whereas late immigrators showed more use of Polish. Both groups displayed proportionally more numerous autobiographical retrievals that were reported as coming to them internally in Polish (vs Danish) for the decades prior to immigration and more in Danish (vs Polish) after immigration. We propose a culture- and language-specific shaping of semantic and conceptual stores that underpins autobiographical and world knowledge.

Internal languages of retrieval: the bilingual encoding of memories for the personal past.

In contrast to most research on bilingual memory that focuses on how words in either lexicon are mapped onto memory for objects and concepts, we focus on memory for events in the personal past. Using a word-cue technique in sessions devoted exclusively to one language, we found that older Hispanic immigrants who had come to the United States as adults internally retrieved autobiographical memories in Spanish for events in the country of origin and in English for events in the U.S. These participants were consistently capable of discerning whether a memory had come to them "in words" or not, reflecting the distinction between purely imagistic or conceptual memories and specifically linguistic memories. Via examination of other phenomenological features of these memories (sense of re-living, sensory detail, emotionality, and rehearsal), we conclude that the linguistic/nonlinguistic distinction is fundamental and independent of these other characteristics. Bilinguals encode and retrieve certain autobiographical memories in one or the other language according to the context of encoding, and these linguistic characteristics are stable properties of those memories over time.