Recuperao de Informao B Cap 02 Modeling Latent

Recuperação de Informação B Cap. 02: Modeling (Latent Semantic Indexing & Neural Network Model) 2. 7. 2, 2. 7. 3 September 27, 1999

Latent Semantic Indexing n Classic IR might lead to poor retrieval due to: u unrelated documents might be included in the answer set u relevant documents that do not contain at least one index term are not retrieved u Reasoning: retrieval based on index terms is vague and noisy n n The user information need is more related to concepts and ideas than to index terms A document that shares concepts with another document known to be relevant might be of interest

Latent Semantic Indexing n n The key idea is to map documents and queries into a lower dimensional space (i. e. , composed of higher level concepts which are in fewer number than the index terms) Retrieval in this reduced concept space might be superior to retrieval in the space of index terms

Latent Semantic Indexing n Definitions u Let t be the total number of index terms u Let N be the number of documents u Let (Mij) be a term-document matrix with t rows and N columns u To each element of this matrix is assigned a weight wij associated with the pair [ki, dj] u The weight wij can be based on a tf-idf weighting scheme

Latent Semantic Indexing n The matrix (Mij) can be decomposed into 3 matrices (singular value decomposition) as follows: u (Mij) = (K) (S) (D)t u (K) is the matrix of eigenvectors derived from (M)(M)t u (D)t is the matrix of eigenvectors derived from (M)t(M) u (S) is an r x r diagonal matrix of singular values where F r = min(t, N) that is, the rank of (Mij)

Computing an Example n Let (Mij) be given by the matrix u Compute the matrices (K), (S), and (D)t

Latent Semantic Indexing n n n In the matrix (S), select only the s largest singular values Keep the corresponding columns in (K) and (D)t The resultant matrix is called (M)s and is given by u (M)s = (K)s (S)s (D)st u where s, s < r, is the dimensionality of the concept space n The parameter s should be u large enough to allow fitting the characteristics of the data u small enough to filter out the non-relevant representational details

Latent Ranking n n The user query can be modelled as a pseudodocument in the original (M) matrix Assume the query is modelled as the document numbered 0 in the (M) matrix The matrix (M)st(M)s quantifies the relantionship between any two documents in the reduced concept space The first row of this matrix provides the rank of all the documents with regard to the user query (represented as the document numbered 0)

Conclusions n n Latent semantic indexing provides an interesting conceptualization of the IR problem It allows reducing the complexity of the underline representational framework which might be explored, for instance, with the purpose of interfacing with the user

Neural Network Model n Classic IR: u Terms are used to index documents and queries u Retrieval is based on index term matching n Motivation: u Neural networks are known to be good pattern matchers

Neural Network Model n Neural Networks: u The human brain is composed of billions of neurons u Each neuron can be viewed as a small processing unit u A neuron is stimulated by input signals and emits output signals in reaction u A chain reaction of propagating signals is called a spread activation process u As a result of spread activation, the brain might command the body to take physical reactions

Neural Network Model n A neural network is an oversimplified representation of the neuron interconnections in the human brain: u nodes are processing units u edges are synaptic connections u the strength of a propagating signal is modelled by a weight assigned to each edge u the state of a node is defined by its activation level u depending on its activation level, a node might issue an output signal

Neural Network for IR: n From the work by Wilkinson & Hingston, SIGIR’ 91 Query Terms Document Terms k 1 Documents d 1 ka ka kb kb kc dj dj+1 kc kt d. N

Neural Network for IR n n n Three layers network Signals propagate across the network First level of propagation: u Query terms issue the first signals u These signals propagate accross the network to reach the document nodes n Second level of propagation: u Document nodes might themselves generate new signals which affect the document term nodes u Document term nodes might respond with new signals of their own

Quantifying Signal Propagation n Normalize signal strength (MAX = 1) Query terms emit initial signal equal to 1 Weight associated with an edge from a query term node ki to a document term node ki: u. Wiq = n wiq 2 sqrt ( i wiq ) Weight associated with an edge from a document term node ki to a document node dj: u. Wij = wij 2 sqrt ( i wij )

Quantifying Signal Propagation n After the first level of signal propagation, the activation level of a document node dj is given by: i Wiq Wij = F which n n i wiq wij sqrt ( i wiq 2 ) * sqrt ( i wij 2 ) is exactly the ranking of the Vector model New signals might be exchanged among document term nodes and document nodes in a process analogous to a feedback cycle A minimum threshold should be enforced to avoid spurious signal generation

Conclusions n n n Model provides an interesting formulation of the IR problem Model has not been tested extensively It is not clear the improvements that the model might provide